Novel gluing pressure head detection system

Abstract

The utility model discloses a novel gluing pressure head detection system, which comprises a gluing pressure head, an industrial personal computer, a pinhole camera and an image AI identification system, and the image AI identification system further comprises an image acquisition module, an image processing module, a feature extraction module, a feature comparison module, an image storage module and a result output module, the image acquisition module is a pinhole camera, the pinhole camera is mounted on a gluing pressure head and is used for acquiring gluing conditions of products, the image storage module comprises a sample database and is used for providing data support for compared images, and the result output module is used for outputting comparison results of the images. According to the utility model, the gluing condition in the existing production process is detected in real time through the image AI identification system, so that compared with the traditional manual detection, the detection time is shortened, the labor cost is reduced, and the production progress is accelerated.

Description

Technical Field

[0001] The utility model relates to the technical field of glue coating detection, in particular to a novel glue coating pressure head detection system. Background Art

[0002] In the battery production process, the main purpose of gluing is to fix the battery cells, end plates and side plates, and to bond and assemble them. With the help of mechanical equipment, the gluing trajectory can be set and the gluing quality can be monitored in real time to ensure the consistency of the module. During the gluing process, it is necessary to ensure that the glue is applied evenly and fully to avoid leaks or accumulation. In the existing production process, it is necessary to additionally inspect the gluing area of ​​the battery after the battery is gluing, which increases the process flow and manual inspection costs.

[0003] Therefore, a new glue coating head detection system was designed. By monitoring the uniformity of glue coating in real time and using images for comparison, the glue coating quality can be evaluated when the workpiece is finished coating, thus shortening the output time of a battery module and speeding up the production progress. Utility Model Content

[0004] The purpose of the utility model is to provide a new type of glue pressure head detection system, which detects the glue coating situation in the existing production process through an image AI recognition system, and performs detection through real-time image acquisition and AI model comparison. Compared with traditional manual detection, it shortens the detection time and labor costs and speeds up the production progress.

[0005] The utility model provides the following technical solutions: a novel glue pressure head detection system, including a glue pressure head, an industrial computer, a pinhole camera and an image AI recognition system, wherein the image AI recognition system also includes an image acquisition module, an image processing module, a feature extraction module, a feature comparison module, an image storage module and a result output module;

[0006] The image acquisition module is a pinhole camera, which is installed on the glue coating pressure head and is used to collect the glue coating situation of the product;

[0007] The image processing module is used to process the pixels of the image to improve the quality of the image;

[0008] The feature extraction and representation module is used to convert the image into information that can be understood and processed by a computer;

[0009] The feature comparison module is used to compare the collected and processed glue coating conditions with the good glue coating products in the sample database;

[0010] The image storage module includes a sample database for providing data support for the compared images;

[0011] The result output module is used to output the image comparison result.

[0012] In order to improve the quality of the image, the image processing module includes an image enhancement module, an edge detection module and an object detection module. The image enhancement module is used to improve the image quality, enrich the information, and enhance the image interpretation and recognition effect. The edge detection module is used to identify points in the image with obvious brightness changes and eliminate irrelevant data. The object detection module is used to detect the target object in the image.

[0013] In order to reduce the data dimension of the image, the feature extraction representation module includes a local pattern extraction module, a texture extraction module, a shape extraction module or a color extraction module. By selecting a suitable module to extract the corresponding features, the image data dimension is reduced and the computational complexity is reduced.

[0014] In order to reduce the complexity of calculation, the feature extraction and representation module converts the features extracted from the image into a vector or vector set that can be processed by a computer. The vector representation maps the image information to a high-dimensional feature space for machine learning. The vector is usually represented as a histogram, vector quantization, local binary pattern (LBP), or histogram of directional gradients (HOG).

[0015] In order to enable the system to learn deeply and improve the recognition rate, the image storage module will store the qualified image data in the sample database after comparing the image with the previous sample database, thereby increasing the number of good glue coating images in the sample database.

[0016] In order to ensure uniform glue coating, a plurality of glue coating fine needle grooves are provided on the end surface of the glue coating pressure head.

[0017] In order to connect the pinhole camera to the glue coating head, a connecting ring is sleeved on the surface of the glue coating head, and the pinhole camera is fixed on the connecting ring.

[0018] Compared with the prior art, the beneficial effects achieved by the utility model are:

[0019] (1) The glue coating situation in the existing production process is detected through the image AI recognition system. The detection is carried out through real-time image acquisition and AI model comparison. Compared with traditional manual detection, it shortens the detection time and labor costs and speeds up the production progress;

[0020] (2) By storing each collected image in a sample database, the AI ​​model automatically learns based on the image, thereby increasing the AI ​​recognition rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0022] Figure 1 It is a schematic diagram of the installation structure of the glue pressing head and the pinhole camera of the utility model;

[0023] Figure 2 It is a schematic diagram of the end structure of the glue coating pressure head of the utility model;

[0024] Figure 3 It is a structural schematic diagram of the gluing pressure head detection system of the utility model;

[0025] Figure 4 It is a schematic diagram of the recognition process of the image AI recognition system of the utility model;

[0026] Figure 5 It is a structural schematic diagram of the image processing module of the utility model;

[0027] Figure 6 It is a structural schematic diagram of the feature extraction and representation module of the utility model;

[0028] Figure 7 It is a schematic diagram of the detection process of the glue pressure head detection system of the utility model;

[0029] In the figure: 1. Glue coating head; 11. Glue coating fine needle groove; 2. Pinhole camera; 21. Connecting ring; 3. Industrial computer; 4. Image AI recognition system; 41. Image acquisition module; 42. Image processing module; 421. Image enhancement module; 422. Edge detection module; 423. Object detection module; 43. Feature extraction and representation module; 431. Local pattern extraction module; 432. Texture extraction module; 433. Shape extraction module; 434. Color extraction module; 44. Feature comparison module; 45. Result output module; 46. Image storage module; 47. Sample database. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the utility model to clearly and completely describe the technical solutions in the embodiments of the utility model. Obviously, the described embodiments are only part of the embodiments of the utility model, not all of the embodiments. Based on the embodiments in the utility model, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the utility model. Example

[0031] See also Figure 3 and 4The utility model provides a technical solution: a new type of glue pressure head detection system, including a glue pressure head 1, an industrial computer 3, a pinhole camera 2 and an image AI recognition system 4, the image AI recognition system 4 also includes an image acquisition module 41, an image processing module 42, a feature extraction module 43, a feature comparison module 44, an image storage module 46 and a result output module 45, the image acquisition module 41 is a pinhole camera 2, the pinhole camera 2 is installed on the glue pressure head 1, and is used to collect the glue coating situation of the product, the image processing module 42 is used to process the pixels of the image to improve the quality of the image, and the feature extraction module 43 is used to convert the image into The information is replaced with information that can be understood and processed by the computer. The feature comparison module 44 is used to compare the collected and processed glue coating conditions with the good glue coating products in the sample database 47. The image storage module 46 includes the sample database 47, which is used to provide data support for the compared images. The result output module 45 is used to output the image comparison results. The collected image is compared with the image in the sample database 47 through the image AI recognition system 4, and the AI ​​model is used to confirm whether the glue coating is uniform, which improves the detection efficiency. Compared with the traditional manual detection of glue coating, it saves labor costs, further shortens the production time, and speeds up the production progress.

[0032] like Figure 5 As shown, the image processing module 42 includes an image enhancement module 421, an edge detection module 422 and an object detection module 423. The image enhancement module 421 is used to improve image quality, enrich information, and enhance image interpretation and recognition effects. The edge detection module 422 is used to identify points in the image where brightness changes significantly and remove irrelevant data. The object detection module 423 is used to detect target objects in the image. The image quality is enhanced by the image enhancement module 421, the edge detection module 422, the object detection module 423, etc., and the value and arrangement of the image pixels are analyzed at the same time to extract features, and the image is classified according to the features.

[0033] like Figure 6 As shown, feature extraction and representation are key steps in the field of image processing and computer vision. They convert images into a form that can be understood and processed by computers for subsequent analysis and tasks; the feature extraction and representation module 43 includes a local pattern extraction module 431, a texture extraction module 432, a shape extraction module 433 or a color extraction module 434. By selecting appropriate modules to extract corresponding features, the image data dimension is reduced, the computational complexity is reduced, and it is helpful to identify and distinguish different image categories.

[0034] Feature representation is to convert the features extracted from the image into a vector or a set of vectors. Vector representation maps image information to a high-dimensional feature space for machine learning. Common feature representation methods include histogram, vector quantization, local binary pattern (LBP), and histogram of oriented gradients (HOG).

[0035] In computer vision and image processing tasks, the quality of feature extraction and representation greatly affects the performance of the algorithm. Feature extraction and representation have the following characteristics:

[0036] (1) Distinguishability: It is able to distinguish images of different categories so that samples of different categories are as far away from each other as possible in the feature space.

[0037] (2) Robustness: It is stable against some changes in the image (such as rotation, scaling, and lighting changes) and can maintain feature consistency.

[0038] (3) Low dimensionality: Convert high-dimensional original image information into low-dimensional feature representation to reduce computational complexity.

[0039] (4) Ease of computation: The process of computing feature representations should be efficient so that it can be applied to large-scale image data.

[0040] After comparing the image with the previous sample database 47 , the image storage module 46 stores the detected image data in the sample database 47 , thereby increasing the number of glue-coated images in the sample database 47 .

[0041] The rise of deep learning has changed the way images are classified. Through an end-to-end learning process, convolutional neural networks automatically learn more discriminative features from raw pixel values, avoiding manual feature design. CNN architectures such as VGG, ResNet, and Inception perform well in image classification tasks. This detection system will have errors in the early stages of use, but with the automatic upgrade of AI, that is, the increasing number of good glue coating images in the sample database 47, the AI ​​recognition rate will also increase greatly. It is estimated that the recognition rate can be infinitely close to 100% in one shift of production.

[0042] A plurality of glue fine needle grooves 11 are provided on the end surface of the glue coating pressure head 1. By providing the plurality of glue fine needle grooves 11, the glue discharge can be evenly distributed.

[0043] like Figure 1 and 2As shown, a connecting ring 21 is sleeved on the surface of the glue coating head 1, and a pinhole camera 2 is fixed on the connecting ring 21. During the glue coating process, the pinhole camera 2 is used to transmit the glue coating picture in real time, which is compared with the good glue coating picture stored in the sample database 47. Good products are selected at the moment of coating, which greatly shortens the output time of a battery module and speeds up the production progress.

[0044] like Figure 7 As shown in the figure, the operation process of the detection system is:

[0045] (1) First create an AI model, train the AI ​​model by adding data, and then review, confirm and publish the trained AI model;

[0046] (2) By applying glue to the battery, the image of the glue application is uploaded in real time;

[0047] (3) Compare the uploaded glue-coating image through the AI ​​model to detect whether the glue-coating in the image is a good product, and output the test result;

[0048] (4) Update the AI ​​model data based on the detected images.

[0049] Finally, it should be noted that the above description is only a preferred embodiment of the utility model and is not intended to limit the utility model. Although the utility model is described in detail with reference to the above embodiments, those skilled in the art can still modify the technical solutions recorded in the above embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the utility model shall be included in the protection scope of the utility model.

Claims

1. A new type of glue pressure head detection system, including a glue pressure head, an industrial computer, a pinhole camera and an image AI recognition system, characterized in that: The image AI recognition system also includes an image acquisition module, an image processing module, a feature extraction module, a feature comparison module, an image storage module and a result output module; The image acquisition module is a pinhole camera, which is installed on the glue coating pressure head and is used to collect the glue coating situation of the product; The image processing module is used to process the pixels of the image to improve the quality of the image; The feature extraction and representation module is used to convert the image into information that can be understood and processed by a computer; The feature comparison module is used to compare the collected and processed glue coating conditions with the good glue coating products in the sample database; The image storage module includes a sample database for providing data support for the compared images; The result output module is used to output the image comparison result.

2. A new type of glue pressure head detection system according to claim 1, characterized in that: The image processing module includes an image enhancement module, an edge detection module and an object detection module. The image enhancement is used to improve image quality, enrich information, and enhance image interpretation and recognition effects. The edge detection module is used to identify points in the image with obvious brightness changes and eliminate irrelevant data. The object detection module is used to detect target objects in the image.

3. A new type of glue pressure head detection system according to claim 1, characterized in that: The feature extraction and representation module includes a local pattern extraction module, a texture extraction module, a shape extraction module or a color extraction module. By selecting a suitable module to extract corresponding features, the dimension of image data is reduced and the computational complexity is reduced.

4. A new type of glue pressure head detection system according to claim 3, characterized in that: The feature extraction and representation module converts the features extracted from the image into a vector or vector set that can be processed by a computer; The vector representation maps the image information into a high-dimensional feature space for machine learning. The vector is usually represented as histogram, vector quantization, local binary pattern (LBP), or histogram of oriented gradients (HOG).

5. A new type of glue pressure head detection system according to claim 1, characterized in that: The image storage module will store the detected image data in the sample database after comparing the image with the previous sample database.

6. A new type of glue pressure head detection system according to claim 1, characterized in that: A plurality of glue coating fine needle grooves are arranged on the end surface of the glue coating pressure head.

7. A new type of glue pressure head detection system according to claim 1, characterized in that: A connecting ring is sleeved on the surface of the glue coating pressure head, and the pinhole camera is fixed on the connecting ring.

Images