Description:FORM 2
THE PATENTS ACT, 1970
(39 OF 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(See section 10 and rule 13)
1. Title of the invention: “Image Processing Based Device for Investment Castings to Measure Dimension and Detect & Categorize Surface-Defect”
2. Applicant:
NAME NATIONALITY ADDRESS
1. Marwadi University
2. Nabhan Yousef
3. Amit V Sata Indian
Marwadi University, Rajkot-Morbi Highway, At Gauridad, Rajkot – 360003, Gujarat, India
4. Pinal Kantesariya
5. Philip Gajera Udhyog 4.0 LLP, Plot No G207/1, Lodhika GIDC, Kalawad Rd, Metoda, Lodhika, Gujarat 360021
3. Preamble to the description
COMPLETE SPECIFICATION
The following specification particularly describes the invention and the manner in which it is to be performed:

Field of the Invention:
The present invention relates to the field of machine vision, particularly to a kind of detecting system. The present invention more particularly related to image processing based device for investment castings to measure dimension, detect and categorize surface-defect.
Background of the Invention:
Investment casting is a widely used manufacturing process in which a molten metal is poured into a mold to create complex shapes and designs with high accuracy and surface finish. However, surface defects in investment castings can lead to rejection of the castings and increase production costs. Therefore, it is important to accurately measure the dimensions and detect and categorize surface defects in investment castings.
Investment casting is a manufacturing process that involves creating a metal part by pouring molten metal into a ceramic mold. The process is commonly used to create complex parts with intricate shapes, such as those used in aerospace, automotive, and medical industries. However, the quality of the cast parts can be affected by various factors, including dimensional accuracy and surface defects.
Traditional methods of measurement and defect detection in investment castings involve manual inspection, which is time-consuming, labor-intensive, and prone to human errors. To overcome these challenges, various automated techniques have been developed in recent years, including image processing-based systems.
To address these issues, an image processing-based device has been developed for investment castings to measure dimension, detect, and categorize surface defects. The device uses a camera to capture images of the cast parts and then applies image processing algorithms to analyze the images and identify any defects or deviations from the desired dimensions.
An image processing-based device for investment castings can capture images of the castings using a camera or other imaging device, and then analyze the images using computer algorithms to measure the dimensions, detect defects, and categorize them into different types. The system can also generate reports and alerts to inform the operator of any defects or deviations from the expected dimensions.
Object of the Invention:
The main objective of the present invention is to develop an image processing-based inspection device for investment castings.
Another objective of the present invention is to enable the device to perform dimensional measurements, detect defects, categorize defects, and quantify defects all in a single device.
Yet another objective of the present invention is to reduce inspection time and improve productivity in the investment casting industry.
Yet another objective of the present invention is to ensure reliable storage of inspection results for future reference and analysis.
Yet another objective of the present invention is to eliminate the need for specific skill sets or domain expertise in the operation of the device.
Yet another objective of the present invention is to enable the streaming of inspection results to a server using the internet for remote monitoring and analysis.
Yet another objective of the present invention is to facilitate proper categorization of defects that can lead to appropriate analytics and decision-making in the investment casting industry.
Summary of the Invention:
The present invention is an inspection device for investment castings that uses image processing and artificial intelligence to measure the dimensions, detect and categorize defects, and quantify their size, all on a single platform. The device is user-friendly and eliminates manual recording of inspection results by streaming them to a server using the internet. The device comprises hardware and software, and the software is developed using more than 3500 images of defective and non-defective investment castings from industrial foundries. The images are pre-processed to remove noise, features highlighting defect occurrence are extracted, and the images are trained using artificial intelligence trained algorithms.
The hardware consists of an aluminum structure, opaque side plate, opaque top plate, base, optical high-resolution camera, LEDs to maintain uniform lighting, user interface, high-end mini-computer, and microcontroller for controlling the up-down movement of the camera. The device's software integrates dimensional measurement, defect detection, categorization, and quantification with the user interface of the hardware. The device provides accurate inspection results with minimal manual intervention and requires no specific skill set or domain expertise. The invention saves inspection time, improves productivity, and eliminates the need for skilled manpower and domain experts in investment casting foundries. The invention provides a reliable storage of inspection results, enabling appropriate analytics. The device is suitable for ferrous as well as nonferrous metallic investment castings.
Brief Description of the Drawings:
Figure 1 shows different views of Image Processing Based Device
Nomenclature Parts
1 Aluminium structure
2 Opaque side plate
3 Opaque top plate
4 Base
5 Optical high resolution camera
6 LED to maintain uniform lighting
7 User interface
8 High end mill computer
9 Microcontroller for controlling up-down movement of camera

Figure 2 shows functional Prototype of image processing device
Figure 3 shows user interface of device
Figure 4 shows capturing image of investment casting
Figure 5 shows input related to identification of number
Figure 6 shows defect detection and categorization
Figure 7 shows quantification of defect and dimensional measurement
Figure 8 shows saving inspected results
Detailed Description of the Invention:
The following description relates to a particular manifestation of the present invention. The investment casting is a popular manufacturing process for producing complex parts with high precision and accuracy. However, the quality of investment castings can be compromised by surface defects such as cracks, porosity, and surface roughness. These defects can lead to significant financial losses for manufacturers and negatively impact their reputation in the market. To overcome this challenge, the present invented device has been developed that utilizes advanced image processing algorithms and software to detect and categorize surface defects and measure the dimensions of investment castings with high accuracy. This technology has the potential to revolutionize the investment casting industry and enhance the quality of products while reducing manufacturing costs.
The present invention is a single inspection device capable of providing results related to measurement of dimension, detection of defects, categorization of defects, and quantification of defects for metallic components manufactured using investment casting process. The device is expected to reduce inspection time and improve productivity while ensuring reliable storage of inspection results. It fills the need for a comprehensive inspection solution that integrates software and hardware, which is not yet available in the market.
The present invention of inspection is an integration of hardware and software, where the hardware component plays a crucial role in the success of the inspection device. The proposed inspection device's hardware is illustrated by figure 1, and it comprises several key components. The first component is the aluminium structure (1), which provides the overall frame for the device. The opaque side plate (2) and top plate (3) are also present to prevent external light from entering the device and affecting the image captured by the camera. The base (4) of the device provides stability to ensure that the inspection process is carried out accurately.
The next essential component of the device is the optical high-resolution camera (5), which captures images of the investment casting for analysis. To ensure that the lighting is uniform and consistent, LEDs (6) are included in the device. This ensures that the captured images are of high quality, which is essential for accurate analysis.
The device is also equipped with a user interface (7), which allows operators to control and monitor the inspection process. A high-end mini computer (8) is also included, which is responsible for processing the images captured by the camera (5) and analyzing them for defects. Finally, a microcontroller (9) is used to control the up-down movement of the camera, ensuring that the entire investment casting is captured for analysis.
These components work together to ensure that the inspection process is carried out accurately and efficiently, providing manufacturers with high-quality investment castings.
Software required in inspection device is developed using fundamentals of artificial intelligence as well as image processing. More than 3500 images related to defective as well as non-defective investment castings were collected from industrial investment casting foundries. These images were pre-processed using specific filters (e.g. Gaussian) to remove noises available. Various features highlighting occurrence of defects were extracted from images using several algorithms (e.g., Harris, Otsu, Hough and Canny). The images were then trained using artificial intelligence trained algorithm (e.g., Convolutional Neural Network), and tested for accuracy of inspection.
The specific method related to dimensional measurement was also programmed, and integrated. Entire bundle of programs related to detection, categorization and quantification of defects with dimensional measurement technique is prepared, and then integrated with user interface of hardware shown in figure 1.
User interface for inspection is shown in figure 3. An integration of hardware and software provides inspection results as per followings steps:
1. It is essential to keep investment casting on base of proposed device and limiting size of investment casting for inspection is 300 mm3, and investment casting should not be weighing more than 20 kg. The proposed device facilitates with LED light (6) to keep uniform lighting on investment castings during inspection.
2. Calibration of camera (5) is one of the very important steps in inspection. The calibration of camera (5) usually comprised of two steps: calibration for position of camera (5), and calibration for dimensional measurement. First step ensures proper alignment in the vertical direction (distance between casting placed on base and camera). The proposed inspected device has facility to adjust vertical alignment of camera (5) using linear actuator that further connected with camera for adjustment of up-down movement. This movement can be controlled using user interface. Second step in calibration of camera (5) is actually carried out using reference object of known dimensions. This step matches help in computing number of pixels used by known reference object, and further assist in mapping overall pixels that are used by investment casting placed on base. Calibration of camera (5) is actually one time activity. Once the information related to calibration is stored in computer then it will be utilized further for dimensional measurement.
3. Input related to investment casting that will be inspected. This input can be of any unique identification of investment casting (e.g., heat number, part number, sample number). This input will be provided using user interface (7). This information will further help in retrieving records related to inspection in future.
4. This step involves capturing image of investment casting placed on base on device using interface (7). Once image is captured then it is shown on screen of user interface (7).
5. This step comprises to click button related to dimension available on user interface (7). This provides dimensional measurement of investment casting placed on base of investment casting.
6. Next step is to click button related to defect then device provides result related to presence of specific type of defect, categorization of the detected defect, and quantification of detected defect.
7. By pressing button related to SAVE, device will stream the inspection results to server through internet. This allows the storage of inspection results on server.
The present device is also tested in an industrial investment casting foundry for inspection and use case highlighting functionalities of inspection deviceis illustrated in figures 4-7.
Table 1: Items used in measurement device
No Item Specification
1 T slotted Aluminum extrusion profiles for structure 40 mm x 40 mm
2 3-way joints --
3 Acrylic opaque sheets 6 mm thick
4 Screen LCD 13.3 inch
5 Linear actuator 1500N-
5.7mms v4
100mm-
6 Optical camera IntelRealSenseDepthCameraD435
7 Microcontroller Arduino Nano
8 Mini computer NUC10i7FNH
9 White LED light --

The present invention create a structure using T slotted aluminum extrusion profiles with specific components, including 3-way joints, acrylic opaque sheets, a screen, a linear actuator, an Intel RealSense Depth Camera D435 (5), an Arduino Nano, a NUC10i7FNH mini-computer, and a white LED light (6). The profiles used for this structure should be 40mm x 40mm in size. The structure should also have a 13.3 inch LCD screen for display. The linear actuator should have a capacity of 1500N, a speed of 5.7mms v4, and a length of 100mm. The Intel RealSense Depth Camera (5) D435 will be used for optical imaging, and the Arduino Nano will be used for control. The NUC10i7FNH mini-computer will provide the computational power for the present device.
To create the structure, the T slotted aluminum extrusion profiles (1) will be used along with the 3-way joints to connect them. The acrylic opaque sheets (2,3) will be used to provide a protective cover for the structure. The screen will be used to display information related to the project. The linear actuator will be used for adjusting the position of the camera (5). The Intel RealSense Depth Camera (5) D435 will be used to capture images of the project. The Arduino Nano will be used to control the linear actuator and camera, and the NUC10i7FNH mini-computer will provide the necessary computational power for the project.
Overall, the combination of these components will create a functional structure that meets the requirements of the present device. The structure will be able to capture images using the Intel RealSense Depth Camera (5) D435, and the linear actuator will allow for precise adjustment of the camera's position. The Arduino Nano will control the linear actuator and camera (5), and the NUC10i7FNH mini-computer will provide the necessary computational power to process the images and display the results on the screen. The white LED light (6) will provide additional illumination for the project.
Main embodiment of the present invention, an Image Processing Based Device for Investment Castings to Measure Dimension and Detect & Categorize Surface-Defect comprising of:
a) an aluminium structure (1) having an opaque side plate (2), an opaque top plate (3), and a base (4);
b) an optical high-resolution camera (5) mounted on the base (4) and plurality of LEDs (6) positioned to provide uniform lighting of the investment casting;
c) a user interface (7) for displaying the inspection results;
d) a high-end mini computer (8) for executing software for measurement of dimension, detection of defects, categorization of defects, and quantification of defects;
e) a microcontroller (9) for controlling the up-down movement of the camera (5); and
f) software is based on the fundamentals of artificial intelligence and image processing, including pre-processing filters to remove noise, algorithms to extract features from images, and artificial intelligence algorithms for training and testing the images;
wherein the said device performs dimensional measurement, detects the presence of surface defects, categorizes the defects into specific types, quantifies the size of the defects, detects more than one surface defect, eliminates manual recording of inspection results, and streams inspection results to a server using internet technology.
Another embodiment of the present invention, the optical high-resolution camera (5) is capable of capturing images with a resolution of at least 10 megapixels.
Another embodiment of the present invention, the opaque side plate (2) and opaque top plate (3) are made of 6 mm thick nontransparent acrylic plate to provide light weight structure.
Another embodiment of the present invention, the LEDs (6) used in the device are white LEDs.
Another embodiment of the present invention, the microcontroller (9) used in the device for controlling the up-down movement of the camera (5) is a stepper motor controller.
Another embodiment of the present invention, the mini-computer (8) used in the device is equipped with a Graphics processing unit for faster image processing.
Another embodiment of the present invention, the software used in the device is capable of detecting defects with a minimum size of 0.1 mm.
Another embodiment of the present invention, the software used in the device is capable of categorizing the defects into surface defects and sub-surface defects.
Another embodiment of the present invention, the software used in the device is capable of quantifying the size of the defects with an accuracy of +/- 5%.
Another embodiment of the present invention, the device is capable of inspecting ferrous and non-ferrous metallic investment castings.
The present invention provides single platform capable of measuring dimensions, detecting and categorizing defects, and quantifying their size for investment castings. The device is also capable of streaming inspection results to a server using the internet, eliminating manual intervention in record keeping. The proposed device's has quick inspection capacity, improved productivity, and appropriate analytics due to proper categorization of defects.
The inspection of metallic components manufactured through investment casting requires accurate measurement of dimensions, detection and categorization of defects, and quantification of their sizes. The proposed device is a unique example of the integration of artificial intelligence and machine learning technologies for achieving such accurate results with minimal need for domain expertise or skills. It provides all the necessary functionalities on a single device, which is not currently available in the market, specifically for investment castings. The device also eliminates the need for manual recording of inspection results. Its features include providing overall dimensional measurements, detecting surface defects on investment castings, categorizing the defects into specific types, quantifying their sizes, detecting and categorizing more than one surface defect, requiring minimum manual intervention, and streaming the inspected results to a server using internet technology. Additionally, the device is suitable for both ferrous and non-ferrous metallic investment castings.
The image processing-based inspection device for investment castings described in this invention provides a user-friendly and efficient way to measure dimensions, detect and categorize defects, and quantify their size. This device is designed to replace the need for skilled manpower and domain experts typically required for inspection-related tasks in investment casting foundries. With the ability to stream inspection results to a server, the device eliminates the need for manual data recording, further increasing its efficiency. The use of advanced algorithms and software to analyze images captured by the camera (5) allows for accurate identification of defects and dimensions, leading to improved efficiency and cost-effectiveness in the manufacturing process. The device can also help manufacturers comply with quality standards and regulations, and provide high-quality products that meet customer specifications. , Claims:We claim,
1. An Image Processing Based Device for Investment Castings to Measure Dimension and Detect & Categorize Surface-Defect comprising of:
a) an aluminium structure (1) having an opaque side plate (2), an opaque top plate (3), and a base (4);
b) an optical high-resolution camera (5) mounted on the base (4) and plurality of LEDs (6) positioned to provide uniform lighting of the investment casting;
c) a user interface (7) for displaying the inspection results;
d) a high-end mini computer (8) for executing software for measurement of dimension, detection of defects, categorization of defects, and quantification of defects;
e) a microcontroller (9) for controlling the up-down movement of the camera (5); and
f) software is based on the fundamentals of artificial intelligence and image processing, including pre-processing filters to remove noise, algorithms to extract features from images, and artificial intelligence algorithms for training and testing the images;
wherein the said device performs dimensional measurement, detects the presence of surface defects, categorizes the defects into specific types, quantifies the size of the defects, detects more than one surface defect, eliminates manual recording of inspection results, and streams inspection results to a server using internet technology.
2. The Image Processing Based Device as claimed in claim 1, wherein the optical high-resolution camera (5) is capable of capturing images with a resolution of at least 10 megapixels.
3. The Image Processing Based Device as claimed in claim 1, wherein the opaque side plate (2) and opaque top plate (3) are made of 6 mm thick nontransparent acrylic plate to provide light weight structure.

4. The Image Processing Based Device as claimed in claim 1, wherein the LEDs (6) used in the device are white LEDs.
5. The Image Processing Based Device as claimed in claim 1, wherein the microcontroller (9) used in the device for controlling the up-down movement of the camera (5) is a stepper motor controller.
6. The Image Processing Based Device as claimed in claim 1, wherein the mini-computer (8) used in the device is equipped with a Graphics processing unit for faster image processing.
7. The Image Processing Based Device as claimed in claim 1, wherein the software used in the device is capable of detecting defects with a minimum size of 0.1 mm.
8. The Image Processing Based Device as claimed in claim 1, wherein the software used in the device is capable of categorizing the defects into surface defects and sub-surface defects.
9. The Image Processing Based Device as claimed in claim 1, wherein the software used in the device is capable of quantifying the size of the defects with an accuracy of +/- 5%.
10. The Image Processing Based Device as claimed in claim 1, wherein the device is capable of inspecting ferrous and non-ferrous metallic investment castings.

Dated 25th Apr, 2023

ChothaniPritibahenBipinbhai
Reg. No.: IN/PA-3148
For and on behalf of the applicant