, Description:FORM 2
THE PATENTS ACT 1970
[39 OF 1970]
&
THE PATENTS RULES, 2003

COMPLETE SPECIFICATION
[See section 10; Rule 13]

TITLE: “A METHOD AND SYSTEM FOR DETERMINING QUALITY OF A COAL SAMPLE”

Name and Address of the Applicants:
(1) TATA STEEL LIMITED, Jamshedpur, Jharkhand, India 831001.
(2) INDIAN STATISTICAL INSTITUTE, 203, Barrackpore Trunk Road, Kolkata, India 700108.

Nationality: India

The following specification particularly describes the invention and the manner in which it is to be performed.
TECHNICAL FIELD
The present subject matter is, in general, related to characterization of coal and more particularly, but not exclusively, to a method and system for determining quality of a coal sample based on analysis of microscopic images of the coal sample.

BACKGROUND
The quality of coal is determined by analyzing various characteristics such as composition of the coal, type of the coal and degree of coalification. A clear understanding of the characteristics of the coal blend is important to decide how coking operations will impact a final product. One of the techniques used for studying the coal characteristics is known as ‘coal petrography’. Coal petrography is a microscopic technique that uses optical microscopic images of the coal sample to determine coal characteristics. Particularly, coal petrography involves use of reflected microscopy to determine a maceral composition and rank of the coal by examining polished specimen of minus 18 mesh prepared coal.
Presently, manual coal petrographic process takes about 5 hours to give a complete analysis of a single coal pellet sample. More importantly, the outcome of manual coal petrography is subjective and depends on the expertise and efficiency of the human expert performing the petrography. On the other hand, there is a continued industry-wide need for an accurate and a reliable petrographic technique, which can be used for studying and selecting coal blends without compromising on coke quality. Also, to optimize a parent coal blend formulation, the objective analysis of the coal samples must be performed quickly. Therefore, there is a need for a faster and an accurate petrographic technique.
The information disclosed in this background of the disclosure section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person skilled in the art.

SUMMARY
Disclosed herein is a method for determining quality of a coal sample. The method comprises analyzing, by a quality analysis system, a plurality of microscopic images of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images. The method further comprises assigning a feature label to each of the one or more features using a pretrained feature classification model. The pretrained feature classification model comprises a feature classifier corresponding to each of the one or more features. Upon assigning the feature labels, the method comprises classifying the feature labels assigned to each of the one or more features into one or more maceral classes for determining a maceral volume distribution corresponding to the coal sample. Finally, the method comprises determining the quality of the coal sample based on the maceral volume distribution.

Further, the present disclosure relates to a quality analysis system for determining quality of a coal sample. The quality analysis system comprises a processor and a memory. The memory is communicatively coupled to the processor and stores processor-executable instructions, which on execution, cause the processor to analyze a plurality of microscopic images of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images. Further, the instructions cause the processor to assign a feature label to each of the one or more features using a pretrained feature classification model. The pretrained feature classification model comprises a feature classifier corresponding to each of the one or more features. Thereafter, the instructions cause the processor to classify feature labels assigned to each of the one or more features into one or more maceral classes for determining a maceral volume distribution corresponding to the coal sample. Finally, the instructions cause the processor to determine the quality of the coal sample based on the maceral volume distribution.

The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, explain the disclosed principles. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The same numbers are used throughout the figures to reference like features and components. Some embodiments of system and/or methods in accordance with embodiments of the present subject matter are now described, by way of example only, and regarding the accompanying figures, in which:

FIG. 1 illustrates an exemplary environment for determining quality of a coal sample in accordance with some embodiments of the present disclosure.

FIG. 2 shows a detailed block diagram of a quality analysis system in accordance with some embodiments of the present disclosure.

FIG. 3A shows a flowchart illustrating a method of training and testing a feature classification model in accordance with some embodiments of the present disclosure.

FIG. 3B shows a flowchart illustrating a method of determining quality of a coal sample in accordance with some embodiments of the present disclosure.

FIG. 4A indicates ground truth labels assigned to a microscopic image of a coal sample using manual petrography in accordance with some embodiments of the present disclosure.

FIG. 4B indicates a neighborhood pixel region used for calculating mean variance histogram of a pixel in accordance with some embodiments of the present disclosure.

FIG. 4C shows an exemplary mean variance histogram in accordance with some embodiments of the present disclosure.

FIG. 5 illustrates a method of performing random forest analysis in accordance with some embodiments of the present disclosure.

FIGS. 6a – 6e illustrate a method of classifying maceral composition of a coal sample in accordance with some embodiments of the present disclosure.

FIGS. 7 and 8 illustrate exemplary classification results in accordance with some embodiments of the present disclosure.

FIG. 9 illustrates a block diagram of an exemplary computer system for implementing embodiments consistent with the present disclosure.

It should be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and executed by a computer or processor, whether such computer or processor is explicitly shown.

DETAILED DESCRIPTION
In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.
While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however that it is not intended to limit the disclosure to the specific forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the scope of the disclosure.
The terms “comprises”, “comprising”, “includes”, or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a setup, device, or method that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus proceeded by “comprises… a” does not, without more constraints, preclude the existence of other elements or additional elements in the system or method.
The present disclosure relates to a method and a quality analysis system for determining quality of a coal sample. In an embodiment, the present disclosure proposes using a machine learning based solution for petrographic phase identification of the coal sample. The proposed solution is automatic and provides better classification accuracy than the manual petrography, in a significantly shorter turnaround time. Moreover, the proposed solution involves calculating unique texture-driven features of different petrographic phases from a plurality of microscopic images of the coal sample, thereby making the classification more comprehensive and reliable. The proposed solution uses a supervised random forest based classification to identify petrographic phases automatically from the microscopic images of the coal sample.

In the following detailed description of the embodiments of the disclosure, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be taken in a limiting sense.

FIG. 1 illustrates an exemplary environment for determining quality of a coal sample in accordance with some embodiments of the present disclosure.

In an embodiment, the environment 100 may include, without limiting to, an image repository 103, a quality analysis system 101 and a feature classification model 107. In an implementation, the image repository 103 may be a storage and/or a database that stores a plurality of microscopic images 105 of a coal sample. The image repository 103 may be communicatively attached to an image capturing device, which may be configured for capturing the plurality of microscopic images 105 of the coal sample. In an alternative implementation, the image repository 103 may be configured within the quality analysis system 101.

In an embodiment, the quality analysis system 101 may be a computing device such as, without limiting to, a desktop computer, a server, a laptop and the like, which may be configured for determining quality of the coal sample in accordance with embodiments of the present disclosure. In an embodiment, the quality analysis system 101 may be deployed at a testing laboratory and used by a petrographer for real-time determination of the coal quality.

In an embodiment, the feature classification model 107 may be a machine learning based supervised learning model, which may be trained for identifying and assigning a feature label for each of one or more features present in the plurality of microscopic images 105 of the coal sample. In an embodiment, the feature classification model 107 may be trained using ground truth feature labels 111 generated by an expert, by manually analysing and labelling the plurality of microscopic images 105. In an implementation, the feature classification model 107 may comprise a plurality of feature classifiers 109, each trained for identifying and classifying a distinct feature present in the plurality of microscopic images 105. In an embodiment, the number of feature classifiers 109 embedded in the feature classification model 107 may be decided based on the number of features to be classified and extracted from the plurality of microscopic images 105.

In an embodiment, when the quality of a coal sample has to be determined, a plurality of microscopic images 105 of the coal sample may be captured in real-time and stored in the image repository 103. Subsequently, the plurality of microscopic images 105 may be forwarded to the quality analysis system 101 for a real-time analysis of the plurality of microscopic images 105. The quality analysis system 101 may analyse the plurality of microscopic images 105 and extract one or more features corresponding to each pixel of each of the plurality of microscopic images 105. The one or more features extracted from the plurality of microscopic images 105 may correspond to one or more coal macerals present in the coal sample. As an example, the one or more coal macerals may include, without limiting to, vitrinite, liptinite, inertinite and other minerals present in the coal sample.

In an embodiment, a pixel and/or a group of pixels in a particular region of the microscopic image, that corresponds to one of the coal maceral classes stated above may be extracted and labelled as a single feature. That is, each pixel of the image that corresponds to maceral type ‘vitrinite’ may be extracted and labelled automatically as belonging to the feature type ‘vitrinite’. Similarly, each of the pixels that correspond to the maceral type ‘liptinite’ and ‘inertinite’ may be extracted and labelled automatically as belonging to feature types ‘liptinite’ and ‘inertinite’ respectively. Further, the remainder of pixels that do not correspond to any of the maceral classes may be extracted and labelled as either resins or other minerals as may be appropriate. Thus, in simple words, the quality analysis system 101 analyses each of the plurality of microscopic images 105 and identifies and extracts each of the features that correspond to a recognized coal maceral class. Thereafter, the quality analysis system 101 assigns a suitable feature label for each of the extracted features with the help of the pretrained feature classification model 107. In one embodiment the feature classification model 107 may be used for labelling the pixels. The training, testing and real-time use of the pretrained feature classification model 107 is explained in detail in the subsequent sections of the detailed description.

In an embodiment, upon assigning feature labels 111 to each of the one or more features extracted from the plurality of microscopic images 105, the quality analysis system 101 may classify the assigned feature labels 111 into one or more maceral classes for determining a maceral volume distribution 113 corresponding to the coal sample. In an embodiment, the maceral volume distribution 113 indicates the proportion of coal sample belonging to each of the one or more maceral classes. For example, the maceral volume distribution 113 discloses what percentage of the total volume of the coal sample is constituted by vitrinite, liptinite, inertinite and other materials.
In an embodiment, upon determining the maceral volume distribution 113, the quality analysis system 101 may determine the quality of the coal sample based on the maceral volume distribution 113. As an example, when the overall percentage of total volume of the maceral classes, namely vitrinite, liptinite and inertinite, is higher than a threshold value, the quality of the coal sample may be rated high. On the other hand, if the maceral volume distribution 113 indicates that the proportion of resins or other minerals is higher in the coal sample, then the coal sample may be considered to be of low quality. Thus, the quality analysis system 101 helps in ready determining the quality of the coal sample based on the plurality of microscopic images 105 of the coal sample.

FIG. 2 shows a detailed block diagram of a quality analysis system 101 in accordance with some embodiments of the present disclosure.

In some implementations, the quality analysis system 101 may include an I/O interface 201, a processor 203 and a memory 205. The I/O interface 201 may be communicatively interfaced with an image repository 103 and/or a microscopic image capturing device for receiving a plurality of microscopic images 105 of the coal sample. Further, the I/O interface 201 may be communicatively interfaced with a feature classification model 107 for obtaining pre-stored machine learning model during real-time classification of the plurality of microscopic images 105. The processor 203 may be configured to perform one or more functions of the quality analysis system 101 for determining quality of the coal sample, using the data 207 and the one or more modules 209. The memory 205 may be communicatively coupled to the processor 203 and may store the data 207 and the one or more modules 209.

In an embodiment, the data 207 may include, without limitation, feature labels 111, maceral volume distribution 113 and other data 215. In some implementations, the data 207 may be stored within the memory 205 in the form of various data structures. Additionally, the data 207 may be organized using data models, such as relational or hierarchical data models. The other data 215 may store various temporary data and files generated by one or more modules 209 while performing various functions of the quality analysis system 101. As an example, the other data 215 may include, without limiting to, microscopic images 105 received from the image repository 103 and/or the microscopic image capturing device, classification model, maceral volume distribution 113 and the like.

In an embodiment, the feature labels 111 may be the labels assigned to each of the one or more features identified and extracted from the plurality of microscopic images 105 of the coal sample. As an example, a pixel and/or a group of pixels that correspond to ‘vitrinite’ may be extracted and assigned with a label ‘vitrinite’. In an embodiment, during real-time analysis of the coal sample, the feature labels 111 for the extracted features may be assigned using the feature classification model 107. That is, once a group of related pixels is extracted from the plurality of microscopic images 105, the features such as mean and variance of the intensity of the pixels, is compared against corresponding feature values pre-stored in the feature classification module for determining the best-match feature label that can be selected and assigned for the extracted pixels.

In an embodiment, the maceral volume distribution 113 may indicate the distribution of volume and/or proportion of each of the one or more coal macerals in the coal sample. In an embodiment, the coal volume distribution may be provided to a user of the quality analysis system 101, as an end result of analyzing the plurality of microscopic images 105. As an example, the maceral volume distribution 113 may be provided on a display interface associated with the quality analysis system 101. Further, the maceral volume distribution 113 may be provided in various formats including, but not limited to, a report, a data table and/or a graphical representation such as pie charts. The maceral volume distribution 113 may be used for determining the quality of the coal sample.

In an embodiment, the data 207 may be processed by the one or more modules 209. In some implementations, the one or more modules 209 may be communicatively coupled to the processor 203 for performing one or more functions of the quality analysis system 101. In an implementation, the one or more modules 209 may include, without limiting to, an analysis module 217, a feature classification model 107, a label classifier 221, a quality determination module 223 and other modules 225.

As used herein, the term module refers to an Application Specific Integrated Circuit (ASIC), an electronic circuit, a hardware processor (shared, dedicated, or group) and memory that execute one or more software or firmware programs, a combinational logic circuit, and/or other suitable components that provide the described functionality. In an embodiment, the other modules 225 may be used to perform various miscellaneous functionalities of the quality analysis system 101. It will be appreciated that such one or more modules 209 may be represented as a single module or a combination of different modules.

In an embodiment, the analysis module 217 may be configured for analyzing a plurality of microscopic images 105 of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images 105. In an embodiment, the analysis module 217 may identify and extract the one or more features from the plurality of microscopic images 105 based on variation of contrast or grey intensities observed in the plurality of microscopic images 105. For instance, the coal sample may have maceral components consisting of vitrinite, inertinite, liptinite and mineral phases embedded in a resin matrix. Some of these maceral components, for example vitrinite and inertinite, have distinct appearances compared to the background resin matrix. However, liptinite and mineral matters may have a low contrast and overlapping appearances. Similarly, even the mineral matters and resin based background regions may have similar appearances under controlled illumination for the microscopic images 105. Further, it may be observed that the liptinite fibers may have a thin strand-like shape and may not have any regular morphology. In an embodiment, the analysis module 217 may take into consideration the aforesaid variations in the appearance of the maceral components to sufficiently differentiate the presence of vitrinite, liptinite, inertinite, mineral matters and the background resin matrix in the plurality of microscopic images 105. Thus, given a microscopic image of the coal sample, the analysis module 217 shall be capable of identifying and extracting distinct features from the microscopic image. In one embodiment, the feature classification model 107 may be used for identifying and extracting distinct features from the microscopic images.

In an embodiment, the feature classification model 107 may be trained for assigning a feature label to each of the one or more features extracted from the plurality of microscopic images 105. In simple words, the feature classification model 107 may be used for naming each of the one or more extracted features with a feature label selected from the group comprising vitrinite, liptinite, inertinite and other minerals/resin matrix. In an embodiment, the feature classification model 107 may be a supervised machine learning based model, which is trained based on ground truth information obtained from a manual petrography of the coal sample by expert petrographers. Once the feature classification model 107 is trained with sufficient features and corresponding feature labels 111, the trained feature classification model 107 is then used for testing unseen microscopic images 105 of the coal sample for evaluation of maceral components.

In an embodiment, the feature classification model 107 may involve use of an ensemble classification, wherein instead of utilizing a single classifier, the trained feature classification model 107 comprises collection of multiple feature classifiers 109. Each of these feature classifiers 109 may be identical in structure, but independent in operation. That is, each of the feature classifiers 109 are trained on separate section of the input data, to ensure that each of the feature classifiers 109 are trained for classification of different features. For example, the feature classification model 107 may be embedded with 4 feature classifiers 109, each of which are trained for classifying the 4 maceral classes namely, vitrinite, liptinite, inertinite and other minerals. The aspect of having an ensemble of multiple feature classifiers 109, each trained for classifying different features, may be compared with the process of having multiple experts simultaneously looking at the microscopic images 105 for parallelly classifying the distinct features from the microscopic images 105. It is as if each one of the experts are generating a class probability of each pixel of the microscopic image belonging to vitrinite, liptinite, inertinite, mineral and resin based background. Finally, the class specific inferences obtained from each of the feature classifiers 109 may be fused according to a standard ensemble approach for completing classification of the plurality of microscopic images 105.

In an embodiment, one of the ensemble approach used for fusing the classification results of each of the one or more feature classifiers 109 may be random forest classifier technique. FIG. 5 provides an exemplary representation of the random forest classifier. As shown in FIG. 5, the random forest may consist of many binary decision trees. A winner feature out of a randomly selected subset of the features divides the pixels into two groups at each node of the binary decision tree. The winner feature may be selected based on a criterion that best divides the dataset into two groups. In an embodiment, the specific advantage of using the random forest classifier may be in terms of its capability to handle data imbalances in different classes of features corresponding to the ground truth labels. Due to expert dependent ground truth labels, there may be significant variations in the number of pixels in belonging to different maceral classes. Therefore, it may be expected that the random forest classifier will perform better in such scenarios. Moreover, because of its parallel architecture, the random forest classifier may be fast and suits very well for real-time analysis of the coal samples.

In an embodiment, the label classifier 221 may be configured for classifying the feature labels 111 assigned to each of the one or more features into one or more maceral classes. In an embodiment, the label classifier 221 may employ a hierarchical approach for classifying the labels based on colour variations in the extracted features. As an example, the dominant colours for both background/resin matrix and mineral regions may be blackish. Based on this observation, the random forest classifier may be modified to a 4-class problem eliminating background class from the training scheme. Further, in the hierarchical approach, the background class may be first separated from the maceral component classes. Next, the classification may be performed only on the regions representing the maceral component classes. Consequently, instead of classifying all the pixels in the image, only the pixels other than the background classes may be classified. As an additional advantage, the execution time for the test module may also be reduced. The hierarchical classification process may be summarized into following steps:

1) Firstly, each of the plurality of microscopic images 105 may be binarized using a predetermined technique such as Otsu’s thresholding. After binarization, the background regions may be marked black and the remaining maceral component regions may be marked white as illustrated in figures – FIG. 6(a) and FIG. 6(b). In an embodiment, the white regions may represent vitrinite, inertinite, liptinite and mineral matters.

2) Small components that may be floating on the resin regions may be replaced as background regions. This may be performed by identifying and eliminating the connected components from the regions. In an embodiment, the connected component may be a component/region, which contains less than a pre-specified number of pixels. Once the small connected components have been eliminated, an image, in which white regions correspond to the maceral components and black regions correspond to the background regions may be obtained as shown in FIG. 6(c).

3) Subsequently, the original microscopic image may be masked using the binarized image obtained in step (1) above to get a resultant masked image as shown in FIG. 6(d). In the masked image, the black pixels may correspond to background pixels and the non-black pixels may correspond to coal components, which need to be classified using the random forest based classifier described earlier. The 4-class classified image corresponding to the masked image of FIG. 6(d) may be as shown in FIG. 6(e).

In an embodiment, the quality determination module 223 (shown in FIG. 2) may be configured for determining the quality of the coal sample based on the maceral volume distribution 113.

FIG. 3A shows a flowchart illustrating a method 300A of training and testing a feature classification model 107 in accordance with some embodiments of the present disclosure.

In an embodiment, at block 301, the quality analysis system 101 may engage and/or instruct the microscopic image capturing device to capture a plurality of microscopic images 105 of the coal sample. At block 303, the quality analysis system 101 may decide to either train the feature classification model 107 with the microscopic images 105 or test/operate the feature classification model 107 on the microscopic images 105. That is, in the training phase, the quality analysis system 101 may proceed to train the feature classification model 107 with a predetermined number of the plurality of microscopic images 105. Alternatively, during the testing phase and/or during real-time analysis of the coal sample, the quality analysis system 101 may activate the feature classification model 107 for analyzing the microscopic images 105, as shown in block 311.

Further, at block 305, that is during the training, ground truth labels may be prepared for the predetermined number of the microscopic images 105 selected for training. In an embodiment, if there are 500 microscopic images 105, then a total of 20-25 images out of the 500 images may be analyzed manually by a petrography expert for preparing the ground-truth labels for the training purpose. The remaining images may be used for testing the trained feature classification module. FIG. 4A(a) shows an exemplary microscopic image in which some portions of the maceral constituents are marked as ground truth labels. In an embodiment, when there may be confusions in labelling a pixel class, a second expert may be considered for unambiguous labelling. Further, a pixel may be rejected for the ground truth analysis, in case if both the experts performing the analysis differ in deciding the class label for the pixel. FIG. 4A(b) shows an exemplary classified microscopic image, in which macerals components such as vitrinite, liptinite, inertinite, minerals and other resin based background regions are represented by different colors - red, magenta, green, cyan and blue respectively.

In an embodiment, after labelling each of the one or more features in the microscopic images 105 selected for training, the quality analysis system 101, at block 307, may extract each of the one or more features identified. Further, at block 309, each of the one or more extracted features and their corresponding feature labels 111 may be fed to the feature classification model 107 for training the feature classification model 107. After successful training, the feature classification model 107 shall be able to identify and label the features from the unseen and/or fresh microscopic images 105 of the coal sample.

Further, at block 311, during testing of the feature classification model 107, the quality analysis system 101 may test the already trained feature classification model 107 by running the feature classification model 107 on the remaining portion of the microscopic images 105 that were not used for training the feature classification model 107 previously, as shown in block 313. In this phase, the pretrained feature classification model 107 may automatically identify and extract each of the one or more features from the plurality of microscopic images 105. Subsequently, the pretrained feature classification model 107 may also assign appropriate feature labels 111 to each of the one or more extracted features. Finally, as indicated in block 315, the one or more features and the corresponding feature labels 111 may be classified for determining the maceral volume distribution 113 of the coal sample. In an embodiment, if the feature classification model 107 is not sufficiently trained and/or not suitable for real-time testing, the quality analysis system 101 may, as shown in block 317, suspend the feature classification model 107 and ensure further training of the feature classification model 107.

FIG. 3B shows a flowchart illustrating a method of determining quality of a coal sample in accordance with some embodiments of the present disclosure.

As illustrated in FIG. 3B, the method 300B may include one or more blocks illustrating a method for determining quality of a coal sample using a quality analysis system 101 illustrated in FIG. 1. The method 300B may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform specific functions or implement specific abstract data types.

The order in which the method 300B is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method. Additionally, individual blocks may be deleted from the methods without departing from the scope of the subject matter described herein. Furthermore, the method can be implemented in any suitable hardware, software, firmware, or combination thereof.

At block 321, the method 300B includes analyzing, by the quality analysis system 101, a plurality of microscopic images 105 of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images 105. In an embodiment, the one or more features may correspond to one or more coal macerals constituting the coal sample. As an example, the one or more coal macerals may include, without limiting to, vitrinite, liptinite, inertinite and other minerals. In an embodiment, each of the one or more coal macerals may be associated with a feature label and a maceral class.

At block 323, the method 300B includes assigning, by the quality analysis system 101, a feature label to each of the one or more features using a pretrained feature classification model 107. In an embodiment, the pretrained feature classification model 107 may comprise a feature classifier corresponding to each of the one or more features. In an embodiment, the feature classification model 107 may be trained by performing a ground-truth analysis of a predefined number of the plurality of microscopic images 105 for extracting the one or more features and then labelling each of the one or more features with respective ground-truth feature labels 111. Subsequently, each of the one or more feature classifiers 109 in the feature classification model 107 may be trained with one of the one or more features and corresponding ground-truth feature labels 111.

At block 325, the method 300B includes classifying, by the quality analysis system 101, the feature labels 111 assigned to each of the one or more features into one or more maceral classes for determining a maceral volume distribution 113 corresponding to the coal sample. In an embodiment, the feature labels 111 may be classified by colouring the one or more pixels associated with the feature labels 111 with a predefined colour corresponding to the one or more maceral classes. Further, classifying the feature labels 111 may include segregating the feature labels 111 into at least one of a background class and a coal component class and eliminating the one or more pixels corresponding to the background class prior to determining the maceral volume distribution 113.

At block 327, the method 300B includes determining, by the quality analysis system 101, the quality of the coal sample based on the maceral volume distribution 113. In an embodiment, the maceral volume distribution 113 may indicate a proportion of each of the one or more coal macerals in the coal sample.

In an embodiment, the one or more features may be extracted from the plurality of microscopic images 105 by splitting each of the plurality of microscopic images 105 into a plurality of image regions having a predefined dimension. Further, a mean and a variance of intensity of each pixel in each of the plurality of image regions may be determined for generating a mean-variance histogram corresponding to each pixel. Subsequently, the mean-variance histogram may be used for determining and extracting the feature corresponding to each pixel of the plurality of image regions.

FIG. 4B indicates a neighborhood pixel region used for calculating mean variance histogram of a pixel in accordance with some embodiments of the present disclosure.

In an embodiment, the mean-variance histogram may be used for each pixel in a certain neighborhood. The mean and variance of pixel values for a certain pre-specified neighborhood may be used to account for the fact that there is no texture signatures for different maceral constituents. The brightness variation between vitrinite, inertinite, mineral and background components may be clearly visible. However, the confusion exists between mineral and background/resin classes and between the liptinite and vitrinite components in certain cases. To counter this, the variance of pixel values of a local neighborhood may be used as a good indicator.

Initially, as shown in FIG. 4B, a 31x31 neighborhood around a pixel may be extracted from the image. The 31x31 region may be further divided into blocks of size 3x3 pixels. For each of the 3x3 blocks, i.e., for 9 pixels, the mean and variance of the pixel values may be calculated. Suppose, the lowest and highest mean values are µl and µh respectively within the 31×31 neighborhood. Similarly, suppose the lowest and highest variance are sl and sh respectively within the 31×31 neighborhood. Then the range of mean (µh-µl) may be divided into a equal bins and the range of variance (sh-sl) may be divided into ß equal bins. This provides aß number of mean-variance bins. Further, the mean-variance histogram may be drawn for aß bins based on the mean-variance values of each 3x3 blocks.

FIG. 4C shows an exemplary mean-variance histogram with a=ß=10. For the intended maceral classification, the feature vector of dimension aß may be used for each pixel of the petrographic image. For each image of size r rows, and c columns, the random forest may accept rc number of features, each of dimension aß. Given the feature extraction, following training and test schemes may be designed:
Training scheme:
During the training, the plurality of microscopic images 105 and their corresponding ground truth images marked with regions of vitrinite, inertinite, liptinite, mineral and background classes may be considered. Different classes of the maceral components in the ground truth images may be marked with different colors as shown in FIG. 4A. Further, from the plurality of microscopic images 105, features of the pixels (i.e., mean-variance histogram) may be extracted from the regions marked as different phases. Also, the corresponding feature labels 111, namely vitrinite, inertinite, liptinite may be collected from the ground truth images. Pixels having same color in the ground truth images may be assigned with the same feature label. In an embodiment, the pixels having same color may be assigned to different classes based on expert and/or manual petrographic analysis. Subsequently, the random forest classifier may be trained using the features and the corresponding class labels of the ground truth pixels. After the training, the trained feature classification model 107 may be obtained, which may be used for testing the unseen microscopic images 105.

Testing scheme:
During testing, a test image, that is a fresh microscopic image of the coal sample may be taken and the features such as mean-variance histogram may be extracted from each pixels of the test image. Since the test image is not present during the training, no features of the test images may be used in the training of the feature classification model 107. Thus, the features extracted from the test image may be fed into the feature classification model 107 for real-time analysis of the test image.

During the testing, each pixel may be assigned a probability of the possible feature class labels. Subsequently, the feature class label corresponding to the highest probability may be assigned to the pixel. From all the assigned feature class labels, the fraction and/or proportion of each maceral classes may be calculated. The appearances of the different phases of the coal microstructures may vary in different samples. Therefore, to capture the variability of the phases for feature extraction, the intensity profile of formation of different phases of the coal may be analyzed rather than using any textural features. The key aspect in capturing the variability of data is the random forest classifier.

A comparison between manual and automated petrography approach:
In an embodiment, the manual calculation of the proportion of the different maceral classes in the coal sample may be time consuming and dependent on the efficiency of the expert/petrographer. Also, in the manual maceral analysis, the percentage of the petrographic phases may be estimated by analyzing and doing at least 500 point counts. Whereas, in the machine learning based approach disclosed in the present disclosure, around 300 images of the coal sample may be considered for estimating the percentage of the petrographic phases as an average of result of analyzing those images. That is, for an image having ‘r’ rows and ‘c’ columns of pixels, the machine learning model of the present disclosure analysis a total of ‘rc’ points. Whereas, in the manual petrographic analysis, only one point per image may be analyzed. This evidently indicates the higher accuracy of petrography analysis resulting from the approach of the present disclosure.

Also, the machine learning based approach of the present disclosure may be fast and accurate than the manual petrography. For instance, the proposed method may perform overall classification 24 times faster compared to the manual petrography. Also, experimental results indicate close to 90% accuracy in the classification for a limited data set for certain petrographic phases like vitrinite, inertinite and the like.

Table A below summarizes results from four different datasets each consisting of approximately 460 images, which is compared with the results obtained from the manual petrography.

Sample name Analysis results from manual petrography Analysis results from the proposed automated method
COAL A Vitrinite=75%
Inertinite= 18%
Liptinite= 3%
Mineral = 4% Vitrinite=69%
Inertinite= 16%
Liptinite= 9%
Mineral = 7%
COAL B Vitrinite=47.7%
Inertinite= 47.2%
Liptinite= 0.3%
Mineral = 4.8%
Vitrinite=47.5%
Inertinite= 46.1%
Liptinite= 0.9%
Mineral = 4.5%
Coke=1.1%
COAL C Vitrinite=64%
Inertinite= 33.6%
Liptinite= 0%
Mineral = 2.4% Vitrinite=54.6%
Inertinite= 44.3%
Liptinite= 0%
Mineral = 1.07%
COAL D Vitrinite=83.8%
Inertinite= 8.7%
Liptinite= 4.2%
Mineral = 3.3% Vitrinite=94.4%
Inertinite= 4.2%
Liptinite= 0.4%
Mineral = 1.1%

Table A – comparison of analysis results

FIG. 7(a) and FIG. 8(a) show the exemplary microscopic images 105 in which the maceral components have been marked white and the remaining background and/or resin components have been marked black. FIG. 7(b) and FIG. 8(b) respectively show the feature labelled images corresponding to the microscopic images 105 of FIG. 7(a) and FIG. 8(a), wherein each maceral component within the white region of the microscopic images 105 are classified with different colors.

Computer System
FIG. 9 illustrates a block diagram of an exemplary computer system 900 for implementing embodiments consistent with the present disclosure. In an embodiment, the computer system 900 may be the quality analysis system 101 illustrated in FIG. 1, which may be used for determining quality of the coal sample. The computer system 900 may include a central processing unit (“CPU” or “processor”) 902. The processor 902 may comprise at least one data processor for executing program components for executing user- or system-generated business processes. A user may include a person, a petrographer, a coal mining engineer, an organization or any system/sub-system being operated parallelly to the computer system 900. The processor 902 may include specialized processing units such as integrated system (bus) controllers, memory management control units, floating point units, graphics processing units, digital signal processing units, etc.
The processor 902 may be disposed in communication with one or more input/output (I/O) devices (911 and 912) via I/O interface 901. The I/O interface 901 may employ communication protocols/methods such as, without limitation, audio, analog, digital, stereo, IEEE®-1394, serial bus, Universal Serial Bus (USB), infrared, PS/2, BNC, coaxial, component, composite, Digital Visual Interface (DVI), high-definition multimedia interface (HDMI), Radio Frequency (RF) antennas, S-Video, Video Graphics Array (VGA), IEEE® 802.n /b/g/n/x, Bluetooth, cellular (e.g., Code-Division Multiple Access (CDMA), High-Speed Packet Access (HSPA+), Global System For Mobile Communications (GSM), Long-Term Evolution (LTE) or the like), etc. Using the I/O interface 901, the computer system 900 may communicate with one or more I/O devices 911 and 912.
In some embodiments, the processor 902 may be disposed in communication with a communication network 909 via a network interface 903. The network interface 903 may communicate with the communication network 909. The network interface 903 may employ connection protocols including, without limitation, direct connect, Ethernet (e.g., twisted pair 10/100/1000 Base T), Transmission Control Protocol/Internet Protocol (TCP/IP), token ring, IEEE® 802.11a/b/g/n/x, etc. Using the network interface 903 and the communication network 909, the computer system 900 may communicate with the image repository 103 and/or the microscopic image capturing device (not shown in FIG. 9) for receiving a plurality of microscopic images 105 of the coal sample. Additionally, the computer system 900 may communicate with a pretrained feature classification model 107 for receiving the feature labels 111 corresponding to features extracted from the plurality of microscopic images 105.
In an implementation, the communication network 909 may be implemented as one of the several types of networks, such as intranet or Local Area Network (LAN) and such within the organization. The communication network 909 may either be a dedicated network or a shared network, which represents an association of several types of networks that use a variety of protocols, for example, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol/Internet Protocol (TCP/IP), Wireless Application Protocol (WAP), etc., to communicate with each other. Further, the communication network 909 may include a variety of network devices, including routers, bridges, servers, computing devices, storage devices, etc.
In some embodiments, the processor 902 may be disposed in communication with a memory 905 (e.g., RAM 913, ROM 914, etc. as shown in FIG. 9) via a storage interface 904. The storage interface 904 may connect to memory 905 including, without limitation, memory drives, removable disc drives, etc., employing connection protocols such as Serial Advanced Technology Attachment (SATA), Integrated Drive Electronics (IDE), IEEE-1394, Universal Serial Bus (USB), fiber channel, Small Computer Systems Interface (SCSI), etc. The memory drives may further include a drum, magnetic disc drive, magneto-optical drive, optical drive, Redundant Array of Independent Discs (RAID), solid-state memory devices, solid-state drives, etc.
The memory 905 may store a collection of program or database components, including, without limitation, user/application interface 906, an operating system 907, a web browser 908, and the like. In some embodiments, computer system 900 may store user/application data 906, such as the data, variables, records, etc. as described in this invention. Such databases may be implemented as fault-tolerant, relational, scalable, secure databases such as Oracle® or Sybase®.
The operating system 907 may facilitate resource management and operation of the computer system 900. Examples of operating systems include, without limitation, APPLE® MACINTOSH® OS X®, UNIX®, UNIX-like system distributions (E.G., BERKELEY SOFTWARE DISTRIBUTION® (BSD), FREEBSD®, NETBSD®, OPENBSD, etc.), LINUX® DISTRIBUTIONS (E.G., RED HAT®, UBUNTU®, KUBUNTU®, etc.), IBM® OS/2®, MICROSOFT® WINDOWS® (XP®, VISTA®/7/8, 10 etc.), APPLE® IOS®, GOOGLE TM ANDROID TM, BLACKBERRY® OS , or the like.
The user interface 906 may facilitate display, execution, interaction, manipulation, or operation of program components through textual or graphical facilities. For example, the user interface 906 may provide computer interaction interface elements on a display system operatively connected to the computer system 900, such as cursors, icons, check boxes, menus, scrollers, windows, widgets, and the like. Further, Graphical User Interfaces (GUIs) may be employed, including, without limitation, APPLE® MACINTOSH® operating systems’ Aqua®, IBM® OS/2®, MICROSOFT® WINDOWS® (e.g., Aero, Metro, etc.), web interface libraries (e.g., ActiveX®, JAVA®, JAVASCRIPT®, AJAX, HTML, ADOBE® FLASH®, etc.), or the like.

The web browser 908 may be a hypertext viewing application. Secure web browsing may be provided using Secure Hypertext Transport Protocol (HTTPS), Secure Sockets Layer (SSL), Transport Layer Security (TLS), and the like. The web browsers 908 may utilize facilities such as AJAX, DHTML, ADOBE® FLASH®, JAVASCRIPT®, JAVA®, Application Programming Interfaces (APIs), and the like. Further, the computer system 900 may implement a mail server stored program component. The mail server may utilize facilities such as ASP, ACTIVEX®, ANSI® C++/C#, MICROSOFT®, .NET, CGI SCRIPTS, JAVA®, JAVASCRIPT®, PERL®, PHP, PYTHON®, WEBOBJECTS®, etc. The mail server may utilize communication protocols such as Internet Message Access Protocol (IMAP), Messaging Application Programming Interface (MAPI), MICROSOFT® exchange, Post Office Protocol (POP), Simple Mail Transfer Protocol (SMTP), or the like. In some embodiments, the computer system 900 may implement a mail client stored program component. The mail client may be a mail viewing application, such as APPLE® MAIL, MICROSOFT® ENTOURAGE®, MICROSOFT® OUTLOOK®, MOZILLA® THUNDERBIRD®, and the like.

Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present invention. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., non-transitory. Examples include Random Access Memory (RAM), Read-Only Memory (ROM), volatile memory, nonvolatile memory, hard drives, Compact Disc (CD) ROMs, Digital Video Disc (DVDs), flash drives, disks, and any other known physical storage media.

Advantages of the embodiments of the present disclosure are illustrated herein.
In an embodiment, the present disclosure provides an automated petrography method for determining quality of coal sample. Consequently, the present disclosure eliminates human intervention and associated human errors involved in the petrographic process.

In an embodiment, the method of present disclosure drastically reduces total turnaround time required for analyzing and determining quality of the coal sample, in comparison to the turnaround time required for manual petrography.

In an embodiment, the method of present disclosure enhances accuracy of classification of the coal composition by using a pretrained, machine learning based classification model for automatically identifying the petrographic phases from the microscopic images.

The aforesaid technical advancement and practical application of the disclosed method and system may be attributed to the aspect of ‘assigning’ feature labels 111 to each of the one or more features and ‘classifying’ feature labels 111 assigned to each of the features into one or more maceral classes for determining a maceral volume distribution 113, as disclosed in the independent claims 1 and 10 of the disclosure.

In light of the technical advancements provided by the disclosed method and system, the claimed steps, as discussed above, are not routine, conventional, or well-known aspects in the art, as the claimed steps provide the aforesaid solutions to the technical problems existing in the conventional technologies. Further, the claimed steps clearly bring an improvement in the functioning of the system itself, as the claimed steps provide a technical solution to a technical problem.

The terms "an embodiment", "embodiment", "embodiments", "the embodiment", "the embodiments", "one or more embodiments", "some embodiments", and "one embodiment" mean "one or more (but not all) embodiments of the invention(s)" unless expressly specified otherwise.

The terms "including", "comprising", “having” and variations thereof mean "including but not limited to", unless expressly specified otherwise.

The enumerated listing of items does not imply that any or all the items are mutually exclusive, unless expressly specified otherwise. The terms "a", "an" and "the" mean "one or more", unless expressly specified otherwise.

A description of an embodiment with several components in communication with each other does not imply that all such components are required. On the contrary, a variety of optional components are described to illustrate the wide variety of possible embodiments of the invention.

When a single device or article is described herein, it will be clear that more than one device/article (whether they cooperate) may be used in place of a single device/article. Similarly, where more than one device or article is described herein (whether they cooperate), it will be clear that a single device/article may be used in place of the more than one device or article or a different number of devices/articles may be used instead of the shown number of devices or programs. The functionality and/or the features of a device may be alternatively embodied by one or more other devices which are not explicitly described as having such functionality/features. Thus, other embodiments of the invention need not include the device itself.

Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based here on. Accordingly, the embodiments of the present invention are intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the true scope and spirit being indicated by the following claims.

Referral Numerals:
Reference Number Description
100 Environment
101 Quality analysis system
103 Image repository
105 Microscopic images
107 Feature classification model
109 Feature classifiers
111 Feature labels
113 Maceral volume distribution
201 I/O interface
203 Processor
205 Memory
207 Data
209 Modules
215 Other data
217 Analysis module
221 Label classifier
223 Quality determination module
225 Other modules
900 Exemplary computer system
901 I/O Interface of the exemplary computer system
902 Processor of the exemplary computer system
903 Network interface
904 Storage interface
905 Memory of the exemplary computer system
906 User/Application
907 Operating system
908 Web browser
909 Communication network
911 Input devices
912 Output devices
913 RAM
914 ROM

Claims:WE CLAIM:
1. A method for determining quality of a coal sample, the method comprising:
analysing, by a quality analysis system, a plurality of microscopic images of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images;
assigning, by the quality analysis system, a feature label to each of the one or more features using a pretrained feature classification model, wherein the pretrained feature classification model comprises a feature classifier corresponding to each of the one or more features;
classifying, by the quality analysis system, feature labels assigned to each of the one or more features into one or more maceral classes for determining a maceral volume distribution corresponding to the coal sample; and
determining, by the quality analysis system, the quality of the coal sample based on the maceral volume distribution.

2. The method as claimed in claim 1, wherein the one or more features correspond to one or more coal macerals constituting the coal sample.

3. The method as claimed in claim 2, wherein the one or more coal macerals comprises vitrinite, liptinite, inertinite and other minerals.

4. The method as claimed in claim 2, wherein each of the one or more coal macerals are associated with a feature label and a maceral class.

5. The method as claimed in claim 1, wherein the feature classification model is trained by:
performing a ground-truth analysis of a predefined number of the plurality of microscopic images for extracting the one or more features;
labelling each of the one or more features with respective ground-truth feature labels; and
training each of the one or more feature classifiers in the feature classification model with one of the one or more features and corresponding ground-truth feature labels.

6. The method as claimed in claim 1, wherein extracting the one or more features comprises:
splitting each of the plurality of microscopic images into a plurality of image regions having a predefined dimension;
determining mean and variance of intensity of each pixel in each of the plurality of image regions; and
generating a mean-variance histogram corresponding to each pixel for determining and extracting the feature corresponding to each pixel of the plurality of image regions.

7. The method as claimed in claim 1, wherein classifying the feature labels comprises colouring the one or more pixels associated with the feature labels with a predefined colour corresponding to the one or more maceral classes.

8. The method as claimed in claim 1, wherein classifying the feature labels further comprises:
segregating the feature labels into at least one of a background class and a coal component class; and
eliminating the one or more pixels corresponding to the background class prior to determining the maceral volume distribution.

9. The method as claimed in claim 1, wherein the maceral volume distribution indicates proportion of each of the one or more coal macerals in the coal sample.

10. A quality analysis system for determining quality of a coal sample, the quality analysis system comprising:
a processor; and
a memory, communicatively coupled to the processor, wherein the memory stores processor-executable instructions, which on execution, cause the processor to:
analyse a plurality of microscopic images of the coal sample for extracting one or more features corresponding to each pixel of each of the plurality of microscopic images;
assign a feature label to each of the one or more features using a pretrained feature classification model, wherein the pretrained feature classification model comprises a feature classifier corresponding to each of the one or more features;
classify feature labels assigned to each of the one or more features into one or more maceral classes for determining a maceral volume distribution corresponding to the coal sample; and
determine the quality of the coal sample based on the maceral volume distribution.

11. The quality analysis system as claimed in claim 10, wherein the one or more features correspond to one or more coal macerals constituting the coal sample.

12. The quality analysis system as claimed in claim 11, wherein the one or more coal macerals comprises vitrinite, liptinite, inertinite and other minerals.

13. The quality analysis system as claimed in claim 11, wherein each of the one or more coal macerals are associated with a feature label and a maceral class.

14. The quality analysis system as claimed in claim 10, wherein the feature classification model is trained by:
performing a ground-truth analysis of a predefined number of the plurality of microscopic images for extracting the one or more features;
labelling each of the one or more features with respective ground-truth feature labels; and
training each of the one or more feature classifiers in the feature classification model with one of the one or more features and corresponding ground-truth feature labels.

15. The quality analysis system as claimed in claim 10, wherein the processor extracts the one or more features by:
splitting each of the plurality of microscopic images into a plurality of image regions having a predefined dimension;
determining mean and variance of intensity of each pixel in each of the plurality of image regions; and
generating a mean-variance histogram corresponding to each pixel for determining and extracting the feature corresponding to each pixel of the plurality of image regions.

16. The quality analysis system as claimed in claim 10, wherein classifying the feature labels comprises colouring the one or more pixels associated with the feature labels with a predefined colour corresponding to the one or more maceral classes.

17. The quality analysis system as claimed in claim 10, wherein classifying the feature labels further comprises:
segregating the feature labels into at least one of a background class and a coal component class; and
eliminating the one or more pixels corresponding to the background class prior to determining the maceral volume distribution.

18. The quality analysis system as claimed in claim 10, wherein the maceral volume distribution indicates proportion of each of the one or more coal macerals in the coal sample..