Description:ARTIFICIAL INTELLIGENCE ENABLED METHOD FOR NON-CONTACT MEASUREMENT OF LINEAR AND ANGULAR FEATURES OF OBJECTS USING MACHINE-LEARNING”
FIELD OF INVENTION
[001] The invention is related to the development of a simple method using Machine Learning for the measurement of the linear and angular dimensions accurately.

[002] In particular, the invention is directed to an artificial intelligence enabled method for non-contact measurement of linear and angular features of objects using machine-learning.

BACKGROUND OF INVENTION

[003] Background description includes information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.

[004] Measurement or metrology is an integral part of any manufacturing sequence. Both contact and non-contact measurements are being used for different purposes such as ensuring the dimensions, measurement of distortion or warpage in case of welding, additive manufacturing, etc. Even though contact type measurements are more reliable and accurate, the use of the same can be laborious and time consuming. For example, in weld induced distortion measurement, contact type measurements may be hazardous due to the heat generated during welding. While in case of image analysis, non-contact measurements can be made any number of times in different time interval. This aids in exact pre-set required to nullify the distortion.

[005] Measuring is also affected by the accessibility of the individual parts. In case of reverse engineering, knocking down all the parts for physical measurement may not be feasible in all cases even with laser stylus/probe and is time consuming too. These measurement constraints limit the measureable parts by the time and cost. Non-contact measurements especially, with machine learning will have an edge in many engineering applications including reverse engineering.

[006] When it comes to additive manufacturing, the warpage in each layer can be measured and corrected by line inspection. Physical measurement may not be possible in every additive manufacturing processes. For example, Wire Arc Additive Manufacturing (WAAM) process is a robust but less accurate than other high energy intense processes. If the Z increment can be dynamically reprogrammed based on the feedback given by the image capturing system can save lot of time and material. The warpage can be corrected dynamically by slightly altering the X and Y coordinates based on the feedback from the image capturing system.

[007] Another case is site inspection, that may involve measurement with radiation hazard or in the oxygen deprived environment are physically unapproachable for contact type measurements, where non-contact measurement techniques provide very less lead time with very less risk of measuring.

[008] Non-contact measuring techniques such as machine vision system are being used for inspection purpose. In traditional non-contact imaging measuring techniques, an exact spitting image of the object to be measured is fed to compare the variations. This process is reliable but is feature specific. Every object inspection has to be separately trained for each individual part. Such machine vision based systems employed in the inspection processes, compares the pre-fed perfect sample with the product in the line. In other words, it can only act as ‘GO’ ‘No GO’ gauge but it cannot absolutely measure the dimension of interest. Hence, a new system of measurement which can absolutely measure the dimensions of the object of interest rapidly will be of great use in many engineering applications.

[009] With the advent of Artificial Intelligence (AI), the required part dimension/variation can be measured by suitable training data. The previously written algorithms can be used only for a specific purpose. These special purpose algorithms cannot be used universally.
[0010] The current methodology is first of a kind invention that can be applied for any objects/features without any restrictions.
PRIOR ART
[0011] Reasonable work has been done for the usage of Machine Learning for engineering applications.

[0012] Reference may be made to below known arts:

[0013] The EU patent WO201359599A1 relates to the machine learning algorithm that captures the two images of the feature to be measured and measures the feature by the pre-fed data of the coordinates of the positions of the camera. The accuracy of the measurement depends on the accuracy of the measurement of the positioning of the cameras.

[0014] The US patent US6611787B2 relates to the measurement of the dimensions based on the Doppler effect from the moving image. The algorithm isolates the object of interest to measure the dimensions of the same.

[0015] The US patent US20220107175 is directed to the measurement of two features that are comparable by size. This algorithm merely acts as a functional gauge whether to allow the product to pass through the line.

[0016] The US patent 11481993 talks about the measurement from the digital images that are pre-processed to convert them into orthographic images. This algorithm can measure only the features that are perpendicular to the each other.
[0017] The US patent US20190310115 relates to the measurement of a feature based on the accelerometer sensor readings and the movement of the image based on the Doppler effect.

[0018] All the prior arts mentioned are discussing apparatus for determining the linear dimensions based on the accelerometers, which limits the measurement only for the moving objects or using two cameras separated at a known distance to measure the dimensions of the object of interest.
OBJECTS OF THE INVENTION
[0019] The object of the present invention is to develop a simple methodology using Machine Learning for quick and reliable measurements for different engineering applications.

[0020] Another objective of the present invention is to determine the distortion angle of the welded plate using an RGB image captured post welding.

[0021] Still another objective of the present invention is to determine the localized buckling angle and slope of the welded sheet metal.

[0022] Yet another objective of the present invention is to determine the variations in dimensions after every layer deposition in the 3-D printing/Additive manufacturing.

[0023] One another objective of the present invention is to determine the warpage in additive manufactured component.

[0024] Yet one another objective of the present invention is to effectively reverse engineer a component by measurement of dimensions.

[0025] These and other objects and advantages of the present invention will be apparent to those skilled in the art after a consideration of the following detailed description taken in conjunction with the accompanying drawings in which a preferred form of the present invention is illustrated.
SUMMARY OF INVENTION
[0026] One or more drawbacks of conventional systems and process are overcome, and additional advantages are provided through the apparatus/composition and a method as claimed in the present disclosure. Additional features and advantages are realized through the technicalities of the present disclosure. Other embodiments and aspects of the disclosure are described in detail herein and are considered to be part of the claimed disclosure.’

[0027] The present developed method utilizes the trained data to determine the angle and the linear dimension of the captured feature from the image to measure the dimensional variations/distortion/buckling/warpage of the object of interest by comparing the intensity difference.

[0028] There is provided an artificial intelligence enabled method for non-contact measurement of linear and angular features of objects using machine-learning comprises of:
[0029] executing an artificial neural network (ANN) for measurement of dimensions of the objects to perform the following steps:
- capturing the image and converting it into the binary data,
- detecting the edges found in the image by internal algorithms,
- eliminating the noise and optimization of the area of interest,
-measuring the pixel data between the edges and matching with the pre-given dimension.
[0030] This is a simple method using Machine Learning algorithm for measurement of dimensions including variations such as thermal expansions.
[0031] It is suitable for detecting edge from a normal RGB image without any direct image processing for accurate measurement of dimensions.
[0032] It is suitable for obtaining distortion angle without any direct image processing.
[0033] It is suitable for obtaining local control points, slopes and angles of the buckled sheet without any direct image processing.

[0034] It is suitable for measurement of warpage in additive manufacturing without any direct image processing.
[0035]
[0036] It is suitable for reverse engineering of components without any direct image processing.
[0037] The main inventive features of the invention have been indicated in the principal claim and subsidiary features in the dependent claims.
[0038] Various objects, features, aspects, and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.

[0039] It is to be understood that the aspects and embodiments of the disclosure described above may be used in any combination with each other. Several of the aspects and embodiments may be combined to form a further embodiment of the disclosure.

[0040] 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 ACCOMAPNYING DRAWINGS
[0041] The illustrated embodiments of the subject matter will be best understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of devices, systems, and processes that are consistent with the subject matter as claimed herein, wherein:-
[0042] Figure 1 shows: Edge detection through the captured RGB image from a colour camera, wherein the captured image is fed into the algorithm, and a random combination of slip gauges is used as a training set, which also includes the training set from the data warehouse.
[0043] Figure 2 shows: Fed test data of the distorted fin.
[0044] Figure 3 shows: Profile and deviation observed in linear dimensions.

[0045] The figures depict embodiments of the disclosure for purposes of illustration only. One skilled in the art will readily recognize from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.
DETAILED DESCRIPTION OF THE INVENTION WITH REFERENCE TO THE ACCOMPANYING DRAWINGS
[0046] While the embodiments of the disclosure are subject to various modifications and alternative forms, specific embodiment thereof have been shown by way of example in the figures and will be described below. It should be understood, however, that it is not intended to limit the disclosure to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the scope of the disclosure.

[0047] The present invention makes a disclosure regarding a technology pertaining to an artificial intelligence enabled method for non-contact measurement of linear and angular features of objects using machine-learning.
[0048] In the present invention, a machine learning algorithm is created in tensor flow, which is trained using the images of the random combination of standard slip gauge set. The pre-known values of different slip gauges are fed as a training set. After training the model, it is used for measuring the required dimensions from the captured image such as weld induced distortion, warpage in additive manufacturing using WAAM, etc. to check the reliability of the developed method.

[0049] For measurement of dimensions or variations, initially, an RGB image is stored as an array of matrix comprising of red, blue and green colour and any sudden change in intensity in any of the defined edges can be identified and used for measurement of dimensions or variations in the object of interest. An effectively trained algorithm can be utilized for any dimensional measurement (linear and angular).

[0050] A normal RGB camera is used to capture the image of the plates and sheets before welding for distortion measurement and in case of additive manufacturing, the standard model is used. After welding/additive manufacturing, with the undisturbed setup, an image is captured. The first image or model is used as a base to estimate the extent distortion of the welded plate or variations w.r.t the desired value in additive manufacturing. The obtained values through this method has been verified by using contact type mechanical measurements and found to be matching.

[0051] This developed simple methodology can be of great use in measurement of dimensions or deviations quickly.

[0052] An algorithm using the convolutional neural network is created, which takes image as an input. The network has to be taught to create an internal regression model based on the data. To teach the algorithm, a standard slip gauge set is used. A different random combinations of the slip gauges was captured using an RGB camera, and is stored in .jpg format. Around 500 different combinations is made and the known dimensions are taught to the algorithm. Internally it forms 4 neural network layers as follows:
-The input layer captures the image and converts it into the binary data. The input layer functions by receiving the captured image (using standard RGB camera) and transforming it into binary data. Additionally, it converts bright field into a 256-bit data set ranging from 0 to 255. Both data sets are subsequently transmitted into the next layer internally.
-The second layer detects the edges found in the image by internal algorithms. The second layer processes the input datasets to determine the displacement from the camera lens to the object of interest through the contrast changes. Simultaneously, it identifies the outer edge by analysing the change in brightness values using the modified Sommerfield algorithm. This detected edge is the one that will undergo measurement, whether it involves angles, linear dimensions or circular features.
-The third layer eliminates the noise and optimizes the area of interest. In this layer, the distinct edge of the object of interest is refined by eliminating potential noise. If the detected edge spans more than 3 pixels in length, this layer applies smoothening techniques tailored to the feature being measured. For instance, it considers the innermost pixel for measuring internal diameter, the outermost pixel for measuring external diameter and the mean pixel data for linear dimensions
-The final layer measures the pixel data between the edges and matches with the pre-given dimension. In the concluding stage, the definitive measurement takes place. Drawing from the contrast data and the redefined edge values, this layer precisely computes various parameters. For linear measurements, it determines the distance between edges accurately. Alternatively, when accessing the angular features, it discerns the angle between the object’s edge and a designated reference line. In cases involving circular attributes, such as diameter, this layer effectively gauges the pertinent dimensions. Notably, these measurements adhere rigorously to the foundational principles extrapolated from teachings of the algorithm using images of the standard slip gauge data, ensuring reliability and precision in the assessment process.
The algorithm fine tunes itself with each iteration and creates a regression model to accommodate all the image data and the fed dimensional data.
TEST RESULT
[0053] The data set of the training data is proportional to the accuracy of the model. So, to increase the accuracy of the model, a data set from the cloud warehouse is utilized to train the model. After training the algorithm with hundred thousand training sets, the model is tested for the accuracy. A separate training set data, irrelevant to the teaching set is created and the model is tested and verified and found to be matching.

[0054] To enhance the accuracy of the model, the model is normalized to avoid overfitting. The expected accuracy interval is lowered to 96% of the training set to prevent the overfitting of the data. After 10 iterations, an accuracy of 98% is achieved in the training set and the 96.6% is achieved in case of the test data.

[0055] After testing the sample data, the real world data is tested in the algorithm. Distortion in the thin fin welding and the warpage in the wire arc additive manufacturing is measured using the developed model and cross verified using contact type mechanical measurements. The measurements made using the developed model were found to be accurate and closely matching with the actual values measured mechanically or by profile projector.
ADVANTAGES OF INVENTION
- Quick and reliable measurements for different engineering applications,
-Determination of the distortion angle of the welded plate using an RGB image captured post welding,
- Determination of the localized buckling angle and slope of the welded sheet metal,
- Determination of the variations in dimensions after every layer deposition in the 3-D printing/Additive manufacturing,
- Determination of the warpage in additive manufactured component,
-Effective reverse engineering of a component by measurement of dimensions.
APPLICATION OF INVENTION
The invention will be of great use in the following fields:
i) Reverse engineering to generate the drawing from a 3D part, where the original drawing is not available;
ii) Accurate measurement of dimensions of additive manufactured parts in every layer and also as a whole;
iii) Comparison of the actually achieved dimensional values with the desired dimensional requirements as per the input drawing in additive manufactured components;
iv) Measurement of linear and angular distortion in weldments especially, long welded structures, components like water wall panels, etc.
[0056] Each of the appended claims defines a separate invention, which for infringement purposes is recognized as including equivalents to the various elements or limitations specified in the claims. Depending on the context, all references below to the "invention" may in some cases refer to certain specific embodiments only. In other cases, it will be recognized that references to the "invention" will refer to subject matter recited in one or more, but not necessarily all, of the claims.

[0057] Groupings of alternative elements or embodiments of the invention disclosed herein are not to be construed as limitations. Each group member can be referred to and claimed individually or in any combination with other members of the group or other elements found herein. One or more members of a group can be included in, or deleted from, a group for reasons of convenience and/or patentability. When any such inclusion or deletion occurs, the specification is herein deemed to contain the group as modified thus fulfilling the written description of all groups used in the appended claims.

[0058] It will be understood by those within the art that, in general, terms used herein, and especially in the appended claims (e.g., bodies of the appended claims) are generally intended as “open” terms (e.g., the term “including” should be interpreted as “including but not limited to,” the term “having” should be interpreted as “having at least,” the term “includes” should be interpreted as “includes but is not limited to,” etc.). It will be further understood by those within the art that if a specific number of an introduced claim recitation is intended, such an intent will be explicitly recited in the claim, and in the absence of such recitation no such intent is present. For example, as an aid to understanding, the following appended claims may contain usage of the introductory phrases “at least one” and “one or more” to introduce claim recitations. However, the use of such phrases should not be construed to imply that the introduction of a claim recitation by the indefinite articles “a” or “an” limits any particulars claim containing such introduced claim recitation to inventions containing only one such recitation, even when the same claim includes the introductory phrases “one or more” or “at least one” and indefinite articles such as “a” or “an” (e.g., “a” and/or “an” should typically be interpreted to mean “at least one” or “one or more”); the same holds true for the use of definite articles used to introduce claim recitations. In addition, even if a specific number of an introduced claim recitation is explicitly recited, those skilled in the art will recognize that such recitation should typically be interpreted to mean at least the recited number (e.g., the bare recitation of “two recitations,” without other modifiers, typically means at least two recitations, or two or more recitations). Furthermore, in those instances where a convention analogues to “at least one of A, B and C, etc.” is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., “a system having at least one of A, B and C” would include but not be limited to systems that have A alone, B alone, C alone, A and B together, A and C together, B and C together, and/or A, B, and C together, etc.). In those instances where a convention analogous to “at least one of A, B, or C, etc.” is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., “a system having at least one of A, B, or C” would include but not be limited to systems that have A alone, B alone, C alone, A and B together, A and C together, B and C together, and/or A, B, and C together, etc.). It will be further understood by those within the art that virtually any disjunctive word and/or phrase presenting two or more alternative terms, whether in the description, claims, or drawings, should be understood to contemplate the possibilities of including one of the terms, either of the terms, or both terms. For example, the phrase “A or B” will be understood to include the possibilities of “A” or “B” or “A and B”.

[0059] The above description does not provide specific details of manufacture or design of the various components. Those of skill in the art are familiar with such details, and unless departures from those techniques are set out, techniques, known, related art or later developed designs and materials should be employed. Those in the art are capable of choosing suitable manufacturing and design details.

[0060] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present disclosure. It will be appreciated that several of the above-disclosed and other features and functions, or alternatives thereof, may be combined into other systems or applications. Various presently unforeseen or unanticipated alternatives, modifications, variations, or improvements therein may subsequently be made by those skilled in the art without departing from the scope of the present disclosure as encompassed by the following claims.
[0061]
[0062] The claims, as originally presented and as they may be amended, encompass variations, alternatives, modifications, improvements, equivalents, and substantial equivalents of the embodiments and teachings disclosed herein, including those that are presently unforeseen or unappreciated, and that, for example, may arise from applicants/patentees and others.
[0063]
[0064] 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.
, Claims:We Claim
1. An artificial intelligence enabled method for non-contact measurement of linear and angular features of objects using machine-learning comprises of:
executing an artificial neural network (ANN) for measurement of dimensions of the objects to perform the following steps:
- capturing the image and converting it into the binary data,
- detecting the edges found in the image by internal mechanism,
- eliminating the noise and optimization of the area of interest,
-measuring the pixel data between the edges and matching with the pre-given dimension.
2. The method as claimed in claim 1, wherein the measurement of dimensions includes variations including thermal expansions.
3. The method as claimed in claim 1 or 2, wherein fine tuning is done with each iteration and a regression model is created to accommodate all the image data and fed dimensional data.
4. The method as claimed in claim 1 , wherein the input layer functions by receiving the captured image and transforming it into binary data, in which the input layer converts bright field into a 256-bit data set ranging from 0 to 255 and both data sets are subsequently transmitted into the next layer internally.
5. The method as claimed in claim 1, wherein the second layer processes the input datasets to determine the displacement from the camera lens to the object of interest through the contrast changes and identifies the outer edge by analysing the change in brightness using the modified Sommerfeld quantization.
6. The method as claimed in claim 1, where in the third layer, the distinct edge of the object of interest is refined by eliminating potential noise such as herein described.
7. The method as claimed in claim 1, where in the final layer, the definitive measurement takes place such as herein described.