Anode carbon block quality detection system

Abstract

The utility model provides a kind of anode carbon block quality detection system, comprising: roller conveyor, for conveying anode carbon block to appearance detection area carries out image and video acquisition;Three high-resolution industrial cameras are respectively used to shoot the image of the top, left side, right side of carbon block;Three high-pixel industrial video cameras are respectively used to shoot the continuous video of the front, back, bottom of carbon block;Weighing sensor is used to weigh the carbon block ready to enter appearance detection area;Intelligent analysis control host is used to generate the complete bottom surface of carbon block according to the continuous video of carbon block bottom surface;For generating the pseudo 3D graph of carbon block according to the image and continuous video of carbon block;For inputting pseudo 3D graph into trained defect classification model, identifying the defect of carbon block;Also used to output corresponding weight grade when the weight of carbon block is consistent with corresponding preset threshold value;Code printing equipment is used to spray defect and weight grade mark on carbon block.

Description

Technical Field

[0001] This utility model relates to the field of electrolytic aluminum technology, specifically to an anode carbon block quality inspection system for quality inspection of anode carbon blocks in the anode assembly workshop of an electrolytic aluminum plant. Background Technology

[0002] Currently, in the anode assembly workshop of aluminum plants, defects such as cracks, missing pieces, missing corners, over-oxidation, or undercooked blocks may occur during the production of carbon blocks. If these defective carbon blocks enter the subsequent assembly and casting stage and are assembled into finished anodes, they will flow into the electrolytic cells in the electrolysis workshop, which will have an adverse effect on the electrolytic cells.

[0003] For defects in the appearance of anode carbon blocks, current methods mainly rely on manual visual inspection, which has a false negative rate of >15% and takes >30 seconds per block. Moreover, bottom defects cannot be detected due to obstruction by the roller conveyor, requiring machine shutdown for resolution. 3D scanning is also used for inspection, but it is costly and not suitable for continuous production lines.

[0004] Chinese invention patent application CN116429770A, published on July 14, 2023, discloses a carbon block surface defect detection system. Although it improves the efficiency of carbon block surface defect detection by using image detection, it only detects the upper surface of the carbon block and does not detect the bottom surface of the carbon block, and only outputs 2D defect images.

[0005] Chinese invention patent application CN108627520A, published on October 9, 2018, discloses an online inspection system and method for the appearance quality of heterogeneous solid materials. Although it greatly improves production efficiency and automation by replacing manual inspection with visual inspection, it is characterized by the need to stop the machine for inspection, making it impossible to conduct continuous inspections. Furthermore, it lacks linkage between weighing and coding, resulting in a lack of data closure. Utility Model Content

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an anode carbon block quality detection system.

[0007] This utility model provides an anode carbon block quality inspection system, comprising: Roller conveyors are used to transport anode carbon blocks to the appearance inspection area for image and video acquisition. Three high-resolution industrial cameras are set up in the appearance inspection area to capture images of the top, left, and right sides of the carbon block, and send the captured images to the intelligent analysis and control host. Three high-resolution industrial cameras are set up in the appearance inspection area to capture continuous video of the front, back, and bottom of the carbon block, and send the captured continuous video to the intelligent analysis and control host. The weighing sensor is installed before the entrance of the appearance inspection area to weigh the carbon block that is about to enter the appearance inspection area and send the weight to the intelligent analysis and control host. The intelligent analysis and control host is used to generate a complete bottom image of the carbon block based on continuous video of the bottom surface of the carbon block; to generate a pseudo 3D image of the carbon block based on images of the top, left, right, and bottom surfaces of the carbon block, as well as continuous video of the front and back; to input the obtained pseudo 3D image of the carbon block into a trained defect classification model to identify defects in the carbon block in the pseudo 3D image; and to output the corresponding weight level when the weight of the carbon block meets the corresponding preset threshold. The inkjet printing equipment is located after the exit of the appearance inspection area and is connected to the intelligent analysis and control host. It is used to spray defect and weight grade markings on the carbon block after the intelligent analysis and control host identifies the defects and weight grade of the carbon block.

[0008] Based on the above, the high-resolution industrial camera used to capture continuous video of the bottom surface of the carbon block adopts a 30° angle to capture video of the roller gap of the roller conveyor.

[0009] Based on the above, it also includes six lighting devices for providing supplementary lighting for three high-resolution industrial cameras and three high-pixel industrial cameras, respectively.

[0010] Based on the above, it also includes a shaded room set up in the appearance inspection area, where three high-resolution industrial cameras and three high-pixel industrial cameras are all installed.

[0011] This utility model has substantial features and progress compared to the prior art, specifically: This invention utilizes a camera to capture video of the roller gaps on a roller conveyor at a 30° angle. This video can then be stitched together to generate a complete image of the carbon block's bottom surface, enabling the identification of defects on the carbon block's bottom surface. Furthermore, since the complete image of the carbon block's bottom surface is obtained by capturing video of the roller gaps on the roller conveyor and then stitching the images together, it is not necessary to stop the roller conveyor during the inspection process. Attached Figure Description

[0012] Figure 1 This is a three-dimensional structural diagram of the present invention.

[0013] In the diagram: 1. Shading room; 2. High-resolution industrial camera; 3. Fill light; 4. Fill light; 5. High-resolution industrial camera; 6. Fill light; 7. High-resolution industrial camera; 8. Fill light; 9. High-resolution industrial camera; 10. Fill light; 11. Fill light; 12. Fill light; 13. High-resolution industrial camera; 14. High-resolution industrial camera; 15. Carbon block; 16. Roller gap; 17. Roller conveyor; 18. Carbon block; 19. Intelligent analysis and control host; 20. Inkjet printing equipment; 21. Weighing controller; 22, 23. Weighing sensor. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0015] like Figure 1 As shown, this embodiment proposes an anode carbon block quality inspection system, including a roller conveyor 17, three high-resolution industrial cameras 5, 7, and 13, three high-pixel industrial cameras 2, 9, and 14, weighing sensors 22 and 23, an intelligent analysis and control host 19, a coding and printing device 20, a light-shielding room 1, supplementary lights 3, 4, 6, 8, 10, 11, and 12, and a weighing controller 21.

[0016] Roller conveyor 17 is used to transport anode carbon blocks 15 to the appearance inspection area for image and video acquisition. The appearance inspection area is divided on roller conveyor 17, and a light-shielding room 1 is installed over the appearance inspection area of ​​roller conveyor 17 to prevent the appearance inspection area from being affected by ambient light during inspection.

[0017] Inside the shaded room 1, three high-resolution industrial cameras (5, 7, and 13), three high-pixel industrial cameras (2, 9, and 14), and supplementary lighting (3, 4, 6, 8, 10, 11, and 12) are installed. High-resolution industrial camera 5 is positioned above the roller conveyor to capture images of the top of carbon block 15; high-resolution industrial camera 7 is positioned to the left of the roller conveyor to capture images of the left side of carbon block 15; and high-resolution industrial camera 13 is positioned to the right of the roller conveyor to capture images of the right side of carbon block 15. Images captured by the three high-resolution industrial cameras 5, 7, and 13 are sent to the intelligent analysis and control host 19. The fill lights 6, 8, and 12 correspond to the settings 5, 7, and 13 of the high-resolution industrial camera, respectively; High-resolution industrial camera 2 is positioned above the roller conveyor and facing the front of carbon block 15 to capture video of the front of carbon block 15. High-resolution industrial camera 9 is positioned above the roller conveyor and facing the back of carbon block 15 to capture video of the back of carbon block 15. High-resolution industrial camera 14 is positioned below the roller conveyor and facing the bottom of carbon block 15 to capture video of the bottom of carbon block 15. Specifically, high-resolution industrial camera 14 uses a 30° angle to capture video of the roller gap 16 of the roller conveyor as the video of the bottom of carbon block 15. The continuous video captured by the three high-resolution industrial cameras 2, 9, and 14 is sent to the intelligent analysis and control host 19. The fill lights 3, 10, and 4 correspond to the settings of high-pixel industrial cameras 2, 9, and 14, respectively.

[0018] Weighing sensors 22 and 23 are installed before the entrance of the appearance inspection area to weigh the carbon block 15 that is about to enter the appearance inspection area and send the weighed weight to the intelligent analysis and control host 19.

[0019] The intelligent analysis and control host 19 is used to generate a complete bottom image of the carbon block based on continuous video of the bottom surface of the carbon block; to generate a pseudo 3D image of the carbon block based on images of the top, left, right, and bottom surfaces of the carbon block, as well as continuous video of the front and back; to input the obtained pseudo 3D image of the carbon block into a trained defect classification model to identify defects in the carbon block in the pseudo 3D image; and to output the corresponding weight level when the weight of the carbon block meets the corresponding preset threshold.

[0020] The inkjet printing device 20 is located after the exit of the appearance inspection area and is connected to the intelligent analysis and control host 19. It is used to spray defect and weight grade markings on the carbon block after the intelligent analysis and control host 19 identifies the defects and weight grade of the carbon block.

[0021] The testing process for this utility model is as follows: Step 1: When carbon block 15 reaches the shooting position, use high-resolution industrial cameras 5, 7, and 13 to take pictures of the top, left, and right sides of carbon block 15; Step 2: As the carbon block passes by the shooting position, the high-pixel industrial camera 9 shoots the video behind it, the high-pixel industrial camera 2 shoots the video in front of it, and the high-pixel industrial camera 14 shoots the video at the bottom. Step 3: The intelligent analysis and control host generates a complete bottom surface image of the carbon block based on the continuous video of the bottom surface of the carbon block; Step 4: The intelligent analysis and control host generates a pseudo 3D image of the carbon block based on the images of the top, left, right, and bottom surfaces of the carbon block, as well as continuous video from the front and back. Step 5: The intelligent analysis and control host inputs the obtained pseudo 3D image of the carbon block into the trained defect classification model to identify the defects of the carbon block in the pseudo 3D image; at the same time, it also outputs the corresponding quality level of the carbon block for subsequent carbon block quality sorting. Step Six: After the intelligent analysis and control host 19 identifies the defects and weight grade of the carbon block, when the carbon block 15 arrives at the inkjet printing device, the inkjet printing device starts to spray defect and weight grade markings on the carbon block; the specific defect marking can be 20250612001-B2 (explanation: the first carbon block on June 12, 2025, defect: missing corner on B2 surface).

[0022] It should be noted that: In this embodiment, the method of generating a complete bottom surface image of the carbon block based on continuous video of the bottom surface of the carbon block can be achieved using existing technologies, such as "A Review of Depth Map Extraction Methods for 2D to 3D Image Conversion", authors: Li Kaihong, Jiang Linmei, Gong Yongyi, journal: Journal of Image and Graphics, published in 2014, volume, issue, page: Vol. 19, No. 10, pp. 1393–1406.

[0023] In this embodiment, the method of generating a pseudo-3D image of the carbon block based on images of the top, left, right, and bottom surfaces of the carbon block, as well as continuous videos of the front and back, can also be implemented using existing technologies, such as "2L3: Generating Accurate 3D Models from Defective 2D Images" (Chen Yizheng et al., 2024); published by Association for Computing Machinery (ACM); ISSN 1557-7368 (Online) / ISSN 0730-0301 (Print).

[0024] In this embodiment, those skilled in the art should understand that the training of the defect classification model can be achieved using existing technologies, such as: Shen Xiaohai et al. "Detection of surface defects of aluminum material based on multi-task deep learning". Progress in Laser & Optoelectronics. 2020, 57 (10).

[0025] In this embodiment, the corresponding weight grades include qualified and unqualified.

[0026] The preset threshold for the weight of the carbon block is 5% of the standard carbon block weight. That is, when the weight of the carbon block is greater than 105% of the standard carbon block weight or less than 95% of the standard carbon block weight, the output weight quality level is unqualified; when 95% of the standard carbon block weight is less than or equal to 105% of the standard carbon block weight, the output weight quality level is qualified.

[0027] The system provided in this application has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A quality inspection system for anode carbon blocks, characterized in that, include: Roller conveyors are used to transport anode carbon blocks to the appearance inspection area for image and video acquisition. Three high-resolution industrial cameras are set up in the appearance inspection area to capture images of the top, left, and right sides of the carbon block, and send the captured images to the intelligent analysis and control host. Three high-resolution industrial cameras are set up in the appearance inspection area to capture continuous video of the front, back, and bottom of the carbon block, and send the captured continuous video to the intelligent analysis and control host. The weighing sensor is installed before the entrance of the appearance inspection area to weigh the carbon block that is about to enter the appearance inspection area and send the weight to the intelligent analysis and control host. The intelligent analysis and control host is used to generate a complete bottom image of the carbon block based on continuous video of the bottom surface of the carbon block; to generate a pseudo 3D image of the carbon block based on images of the top, left, right, and bottom surfaces of the carbon block, as well as continuous video of the front and back; to input the obtained pseudo 3D image of the carbon block into a trained defect classification model to identify defects in the carbon block in the pseudo 3D image; and to output the corresponding weight level when the weight of the carbon block meets the corresponding preset threshold. The inkjet printing equipment is located after the exit of the appearance inspection area and is connected to the intelligent analysis and control host. It is used to spray defect and weight grade markings on the carbon block after the intelligent analysis and control host identifies the defects and weight grade of the carbon block.

2. The anode carbon block quality detection system according to claim 1, characterized in that: A high-resolution industrial camera used to capture continuous video of the bottom surface of carbon blocks, employing a 30° angled view to capture video of the roller gaps on the roller conveyor.

3. The anode carbon block quality detection system according to claim 1 or 2, characterized in that: It also includes six lighting devices for providing supplemental lighting for three high-resolution industrial cameras and three high-pixel industrial cameras, respectively.

4. The anode carbon block quality detection system according to claim 3, characterized in that: It also includes a shaded room set up in the appearance inspection area, where three high-resolution industrial cameras and three high-pixel industrial cameras are housed.

Images