Camera stability monitoring method and system for battery visual inspection equipment

Abstract

The invention discloses a camera stability monitoring method and system for battery visual inspection equipment, belongs to the technical field of lithium battery production, and solves the problem that continuous monitoring cannot be performed during lithium battery point inspection. And the accuracy of subsequent size measurement is ensured. Because the size and the position of the standard part in the image are known, camera stability monitoring is converted into the size detection stability problem of the standard part in the image, the actual size of the standard part in the image is calculated, and the measured actual size is compared with the actual size of the known standard part; any deviation in the size of the standard component can be interpreted as a change in the stability of the camera. According to the method, the stability condition of the camera is accurately identified through the size detection algorithm based on the edge gradient, the accuracy of measured data is guaranteed, and the product quality is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of lithium battery production and relates to a camera stability monitoring method and system for a battery visual inspection device. BACKGROUND

[0002] Lithium batteries are the key energy storage solutions in the fields of mobile communication, electric vehicles, and energy storage systems, and accurate measurement of product size in the production process is of decisive significance to ensure product quality. Visual inspection technology plays an important role in this process, not only improving the speed and accuracy of detection, but also greatly reducing labor costs. In visual inspection technology, cameras are mainly used as monitoring devices, and the stability of the camera is the basis for ensuring the accuracy and reliability of the detection data. However, existing lithium battery visual inspection equipment often fails to maintain the stability of the camera in long-term operation or complex environments, resulting in deviations in camera measurement data and affecting the accurate assessment of the state of lithium batteries. Currently, the camera stability monitoring method commonly used on the production line is manual spot checking. Manual spot checking has obvious limitations, not only consuming a large amount of human resources, but also being low in work efficiency. This intermittent checking method and manual operation cannot perform continuous monitoring, directly affecting the real-time nature of the monitoring data, failing to meet the real-time monitoring needs, and potentially leading to potential problems not being discovered and addressed in a timely manner, thereby affecting product quality and the stability of the production process.

[0003] Prior art such as the invention patent with application publication number CN115731165A discloses a detection method for monitoring the camera stability of lithium battery equipment online, by collecting image information of the calibration unit, then determining the region position of the calibration unit and obtaining the edge coordinate data of the calibration unit region through a template matching method based on gray value, obtaining the calculated space data information of the calibration unit region, and calculating the deviation value of the calculated space data information and the actual space data information of the calibration unit. Finally, it is determined whether there is a detection error in the size detection system according to the deviation value. However, the camera cannot perform continuous monitoring in this online automatic spot checking method, directly affecting the real-time nature of the monitoring data. This intermittent checking method and manual operation cannot perform continuous monitoring, directly affecting the real-time nature of the monitoring data, failing to meet the real-time monitoring needs, and potentially leading to potential problems not being discovered and addressed in a timely manner, thereby affecting product quality and the stability of the production process. SUMMARY

[0004] The technical solution of the present application is used to solve the problem that the camera cannot perform continuous monitoring during lithium battery spot checking.

[0005] The present application solves the above technical problems through the following technical solutions:

[0006] A camera stability monitoring method for a battery visual inspection device, comprising the following steps:

[0007] S1, installing a standard part in the camera field of view, and acquiring a standard part image by the camera;

[0008] S2, calibrating the standard part image based on a pixel accuracy algorithm;

[0009] S3, accurately positioning the position of the standard part in the image based on a template matching algorithm;

[0010] S4, measuring the size of the standard part in the image based on an edge gradient-based size detection algorithm, and monitoring the stability of the camera;

[0011] If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; if the detected size of the standard part in the image fluctuates, it indicates that the stability of the camera is affected;

[0012] S5, comparing the measured actual size with the actual size of the known standard part, and any deviation of the size of the standard part is regarded as a change in the stability of the camera, and the stability of the camera is continuously monitored.

[0013] The present application acquires a standard part image by a camera, calibrates the standard part image to determine the correspondence between the actual size and the pixel size, and ensures the accuracy of subsequent size measurement. Since the size and position of the standard part in the image are known, the present application converts camera stability monitoring into a size detection stability problem of the standard part in the image, calculates the actual size of the standard part in the image, compares the measured actual size with the actual size of the known standard part, and any deviation of the size of the standard part can be explained as a change in the stability of the camera. The present application accurately identifies the stability condition of the camera through the edge gradient-based size detection algorithm, ensures the accuracy of the measurement data, and improves the product quality.

[0014] Further, the S2 specifically calibrates the standard part image captured by the camera to establish the correspondence between the actual size and the pixel size of the standard part in the image, and uses the following logic expression:

[0015]

[0016] Wherein, P pixel is the pixel accuracy, D actual is the actual size, and N pixels is the number of pixels.

[0017] Further, the S3 uses the following logic expression to accurately position the position of the standard part in the image:

[0018]

[0019] where I(x, y) is the pixel value at (x, y) in the input image, T(x, y) is the pixel value at (x, y) in the template image, is the average gray value of the input image, is the average gray value of the template image, R(u, v) is the correlation coefficient of the input image and the template image at position (u, v).

[0020] Further, the S4 further comprises the following steps:

[0021] S41, performing median filtering on the standard part image;

[0022] S42, performing gradient projection on the standard part image, including horizontal gradient projection and vertical gradient projection;

[0023] S43, performing gradient traversal and variance filtering on the standard part image;

[0024] S44, fitting the point set of stable edge points of the standard part image;

[0025] S45, calculating the actual size of the standard part in the image.

[0026] Further, the S42 specifically comprises: firstly determining the horizontal and vertical edges of the object in the standard part image, secondly calculating the gradient sum of each row to perform horizontal gradient projection, and then calculating the gradient sum of each column to perform vertical gradient projection, and performing gradient projection on the standard part image by using the following logical expression:

[0027]

[0028] where P horizontal is the total gradient intensity projection of the image in the horizontal direction, P vertical is the total gradient intensity projection of the image in the vertical direction, G x (x, y) is the gradient of position (x, y) in the input image in the x direction, and G y (x, y) is the gradient of position (x, y) in the input image in the y direction.

[0029] Further, the S43 specifically comprises: firstly traversing the image gradient value to find the edge point whose gradient value exceeds a preset threshold, secondly calculating the variance in the neighborhood of the edge point, and if the variance is less than a set threshold, considering that the point is a stable edge point; if the variance is greater than the set threshold, considering that the point is an unstable edge point, removing the unstable edge point, and calculating the variance in the neighborhood of the edge point by using the following logic:

[0030]

[0031] Wherein, sigma 2 is variance, N is the number of points in the neighborhood, x i is the value of the i-th point, mu is the average value of points in the neighborhood.

[0032] Further, the S44 is specifically: using least square fitting for stable edge points to perform standard part edge straight line fitting, the straight line fitting mode is y=ax+b, wherein y represents, x represents, a represents the slope, and b represents the intercept, and the slope a and the intercept b are solved by using the following logic:

[0033]

[0034] Wherein, x i and y i are the horizontal coordinate and vertical coordinate values representing the i-th edge point.

[0035] Further, the S45 uses the following logic to calculate the actual size of the standard part in the image:

[0036]

[0037] Wherein, (x, y) is a point coordinate selected on the standard part edge straight line, N' pixels is the pixel number of the standard part in the image.

[0038] Further, the S5 uses the following logic to calculate the camera stability Stability Change:

[0039] Stability Change=(N' pixels *P pixel )-D actual

[0040] Wherein, P pixel is the pixel accuracy, D actual is the actual size, N' pixels is the pixel number of the standard part in the image, and if Stability Change is more close to 0, the better the camera stability is.

[0041] The application also provides a camera stability monitoring system for a battery visual detection device, comprising:

[0042] A hardware device module is used to install a standard part in the field of view of a camera, and the standard part image is collected by the camera;

[0043] An image calibration module is used to calibrate the standard part image based on a pixel accuracy algorithm;

[0044] A matching positioning module is used to accurately position the position of the standard part in the image based on a template matching algorithm.

[0045] A stability monitoring module measures the size of the standard part in the image based on the edge gradient size detection algorithm, and monitors the stability of the camera;

[0046] If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; if the detected size of the standard part in the image fluctuates, it indicates that the stability of the camera is affected;

[0047] A stability evaluation module compares the measured actual size with the actual size of the known standard part, and any deviation in the size of the standard part is considered as a change in the stability of the camera, and the stability of the camera is continuously monitored.

[0048] The advantages of the present application are:

[0049] (1) The present application first acquires the image of the standard part by the camera, calibrates the standard part image to determine the correspondence between the actual size and the pixel size, ensures the accuracy of subsequent size measurement, and accurately positions the position of the standard part in the image by using the normalized correlation coefficient template matching technology.

[0050] Secondly, since the size and position of the standard part in the image are known, the present application converts the camera stability monitoring into the size detection stability problem of the standard part in the image, and calculates the actual size of the standard part in the image after the standard part image is subjected to median filtering, gradient projection, gradient traversal, variance filtering and point set fitting in sequence, compares the measured actual size with the actual size of the known standard part, and any deviation in the size of the standard part can be explained as a change in the stability of the camera. The present application accurately identifies the stability condition of the camera by using the edge gradient-based size detection algorithm, ensures the accuracy of the measurement data, and improves the product quality.

[0051] (2) The present application realizes 24-hour uninterrupted real-time monitoring of the camera state by deploying monitoring hardware and developing efficient detection algorithms, introducing a real-time online monitoring system. The need for manual intervention is eliminated, not only saving human resources, but also overcoming the problem of insufficient timeliness of traditional manual point inspection, thereby significantly improving the overall production efficiency. Realize online real-time monitoring of the stability of the camera, thereby completely abandoning the traditional manual point inspection mode, ensuring the accuracy and timeliness of the monitoring data, and significantly improving the production efficiency and product quality control standard.

[0052] (3) The stability monitoring method of the present application is simple, easy to operate, and has the advantages of low cost maintenance, effectively realizing the automatic real-time monitoring of the stability of the camera. Replaces the traditional manual inspection method, not only greatly reduces the labor cost, but also avoids the problems caused by the insufficient timeliness of manual inspection. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 is a flow chart of a camera stability monitoring method for a battery visual inspection device according to an embodiment of the present application;

[0054] Figure 2 is an example diagram of a standard hardware physical display according to an embodiment of the present application;

[0055] Figure 3 is an example diagram of image calibration display according to an embodiment of the present application;

[0056] Figure 4 is an example diagram of image matching positioning display according to an embodiment of the present application;

[0057] Figure 5 is an example diagram of image stability monitoring display according to an embodiment of the present application;

[0058] Figure 6 is an example diagram of continuous image stability monitoring display according to an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in a clear and complete manner with reference to the drawings accompanying the present application and specific embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0060] The technical solutions of the present application will be further described below with reference to the drawings accompanying the present application and specific embodiments:

[0061] Embodiment One

[0062] In this embodiment, the monitoring device uses a high-resolution industrial camera, uses high-precision standard hardware as the monitoring object, manually deploys the standard hardware and fixes it in the camera field of view, which is used for camera stability monitoring, captures images using a high-resolution camera, and combines image processing algorithms to accurately monitor the stability of the camera. Among them, the industrial camera can be selected as an industrial line scanning camera or a plane array camera.

[0063] As shown in Figure 1 , specifically, a camera stability monitoring method for a battery visual inspection device is disclosed, comprising the following steps:

[0064] S1, hardware device, install standard parts in the camera field of view, and capture standard part images through the camera.

[0065] In this embodiment, the standard part is a customized sample with known actual dimensions, which can be 5mm*5mm. The accuracy of the standard part is set to ±0.01mm. A physical image of the standard part is captured using a camera, as shown below. Figure 2 As shown.

[0066] S2, Image calibration, calibrates standard part images based on pixel precision algorithms.

[0067] In this embodiment, the standard part image captured by the camera is calibrated to establish the correspondence between the actual size of the standard part in the image and the pixel size, ensuring the accuracy of subsequent size measurements. The following logic represents the calibration of the standard part image based on a pixel accuracy algorithm:

[0068]

[0069] Among them, P pixel It refers to pixel precision, D actual This is the actual size, N pixels This refers to the number of pixels. In this embodiment, the standard part is image calibrated based on its actual size and the number of camera pixels. The calculated pixel accuracy represents the correspondence between the actual size of the standard part and the pixel size. Figure 3 The image obtained after calibration has a pixel precision of P. pixel =0.083171.

[0070] S3, Matching and Localization: Based on template matching algorithm, this method accurately locates the position of the standard part in the image, using the following logical representation:

[0071]

[0072] Where I(x,y) is the pixel value at (x,y) in the input image, and T(x,y) is the pixel value at (x,y) in the template image. It is the average gray value of the input image. R(u,v) is the average gray value of the template image, and R(u,v) is the correlation coefficient between the input image and the template image at position (u,v). This embodiment uses normalized correlation coefficient template matching technology, such as... Figure 4 The image shows the location of the standard part within the dashed box that was matched and located. The template image can be any selected standard image, and the average gray value of the input image can be the sum of the gray values ​​of all pixels in the input image divided by the total number of pixels.

[0073] S4, stability monitoring, measures the size of a standard part in an image based on an edge gradient-based size detection algorithm to monitor the stability of the camera.

[0074] In the present embodiment, since the size and position of the standard part are known, any detected change in the size of the standard part can be interpreted as a change in the camera stability. If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; on the contrary, if the detected size of the standard part in the image fluctuates, it indicates that the camera stability is affected. The present embodiment converts the camera stability monitoring into a size detection stability problem of the standard part in the image, and the size detection is performed based on an edge gradient-based size detection algorithm, and further, the S4 further comprises the following steps:

[0075] S41, median filtering is performed on the standard part image, for each pixel point in the image, the pixel values in its neighborhood are sorted, and the median value is taken as the filtered value of the pixel point, which is used to reduce image noise and improve the accuracy of edge detection. In the present embodiment, the standard part image is median filtered using the following logic expression:

[0076] I filtered = median[I(x-1,y-1),I(x-1,y),I(x-1,y+1),I(x,y-1),I(x,y),I(x,y+1),I(x+1,y

[0077] -1),I(x+1,y),I(x+1,y+1)]

[0078] Where I(x,y) is the pixel value at position (x,y) in the input image, I filtered is the filtered pixel value, and median[·] is the median operation.

[0079] S42, gradient projection is performed on the standard part image, including horizontal gradient projection and vertical gradient projection, specifically: first, determine the horizontal and vertical edges of the object in the standard part image, second, calculate the gradient sum of each row in the entire standard part image region, perform horizontal gradient projection, then calculate the gradient sum of each column, perform vertical gradient projection. In the present embodiment, the standard part image is gradient projected using the following logic expression:

[0080]

[0081] Where P horizontal is the total gradient intensity projection of the image in the horizontal direction, P vertical is the total gradient intensity projection of the image in the vertical direction, G x (x,y) is the gradient of position (x,y) in the input image in the x direction, which is represented by central difference G x (x,y) = I(x+1,y) - I(x-1,y), G y(x,y) is the gradient of the position (x,y) in the input image in the y direction, expressed by central difference G y (x,y) = I(x,y+1) - I(x,y-1).

[0082] S43, gradient traversal and variance filtering are performed on the standard part image, specifically: first, the gradient values of the entire standard part region image are traversed to find edge points with gradient values exceeding a preset threshold, second, the variance in the neighborhood of the edge points is calculated, if the variance is less than a set threshold, the point is considered to be a stable edge point; if the variance is greater than the set threshold, the point is considered to be an unstable edge point, and the unstable edge points are removed. In this embodiment, the variance in the neighborhood of the edge points is calculated using the following logic:

[0083]

[0084] where σ 2 is the variance, N is the number of points in the neighborhood, x i is the value of the i-th point, and μ is the average value of the points in the neighborhood.

[0085] In this embodiment, the edge point specifically refers to a pixel point of the standard part edge, the preset threshold is the degree of change in the gray scale around the edge point, and the reference setting is 50-150, which is dynamically adjusted according to the actual different image effects.

[0086] S44, the point set of the stable edge points of the standard part image is fitted, specifically: the least square method is used to fit the stable edge points to perform standard part edge straight line fitting, the straight line fitting method is y = ax + b, where a represents the slope and b represents the intercept, and the slope a and the intercept b are solved using the following logic:

[0087]

[0088] where x i and y i are the horizontal and vertical coordinate values of the i-th edge point.

[0089] S45, the actual size of the standard part in the image is calculated, and the following logic is used to represent it:

[0090]

[0091] where (x,y) is a point coordinate selected from the standard part edge straight line, N' pixels is the number of pixels of the standard part in the image.

[0092] S5, stability evaluation, comparing the measured actual size with the actual size of the known standard part, any deviation of the standard part size can be explained as the change of camera stability, and the camera stability is continuously monitored, and the camera stability Stability Change is calculated by using the following logic:

[0093] Stability Change=(N' pixels *P pixel )-D actual

[0094] Wherein, P pixel is the pixel accuracy, D actual is the actual size, N' pixels is the number of pixels of the standard part in the image, if Stability Change is closer to 0, the better the camera stability, such as Figure 5 The standard part stability deviation value in the dotted frame is 0.050mm, which proves that the measured standard part size has small deviation from the actual size.

[0095] The continuous monitoring of the camera stability is specifically: first, image calibration is performed, steps S3-S4 are performed on the initial standard part image captured by the camera, the calculated pixel number value is assigned to N pixels , and then the pixel accuracy P pixel is calculated; then the standard part image captured by the camera is continuously monitored, steps S3-S5 are cycled, N' pixels and Stability Change are continuously and real-timely calculated and recorded. For example, Figure 6 The data of the continuous monitoring of the camera stability is shown in the figure, the camera stability has been maintained between 0-0.1 during the continuous monitoring time, which proves that the camera has good stability and supports continuous monitoring.

[0096] The application also provides a camera stability monitoring system for a battery visual detection device, comprising:

[0097] A hardware device module is used to install the standard part in the camera field of view, and the standard part image is captured by the camera;

[0098] An image calibration module is used to calibrate the standard part image based on the pixel accuracy algorithm;

[0099] A matching positioning module is used to accurately position the position of the standard part in the image based on the template matching algorithm;

[0100] A stability monitoring module is used to measure the size of the standard part in the image based on the edge gradient size detection algorithm, and the camera is monitored for stability;

[0101] If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; if the detected size of the standard part in the image fluctuates, it indicates that the stability of the camera is affected;

[0102] The stability evaluation module is configured to compare the measured actual size with the actual size of the known standard part, and any deviation of the standard part size is taken as a change in the stability of the camera, and the stability of the camera is continuously monitored.

[0103] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit the same; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A camera stability monitoring method for a battery vision inspection apparatus, characterized by, The method comprises the following steps: S1, installing a standard part in the field of view of a camera, and capturing an image of the standard part by the camera; S2, calibrating the image of the standard part based on a pixel accuracy algorithm; S3, accurately positioning the position of the standard part in the image based on a template matching algorithm; S4, measuring the size of the standard part in the image based on an edge gradient size detection algorithm, and monitoring the stability of the camera; If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; If the detected size of the standard part in the image fluctuates, it indicates that the stability of the camera is affected; S5, comparing the actual size measured with the actual size of the known standard part, and any deviation of the size of the standard part is taken as a change in the stability of the camera, and the stability of the camera is continuously monitored.

2. The camera stability monitoring method for battery vision inspection apparatus according to claim 1, wherein, The S2 specifically comprises: calibrating the image of the standard part captured by the camera, establishing the corresponding relationship between the actual size of the standard part in the image and the pixel size, and using the following logic to represent: where P oixel is the pixel precision, D actual is the actual size, and N pixels is the number of pixels.

3. The camera stability monitoring method for battery vision inspection apparatus according to claim 1, wherein, The S3 uses the following logic to accurately position the position of the standard part in the image: where I(x, y) is the pixel value at (x, y) in the input image, T(x, y) is the pixel value at (x, y) in the template image, is the average gray value of the input image, is the average gray value of the template image, and R(u, v) is the correlation coefficient of the input image and the template image at position (u, v).

4. The camera stability monitoring method for battery vision inspection apparatus according to claim 1, wherein, The S4 further comprises the following steps: S41, performing median filtering on the image of the standard part; S42, performing gradient projection on the image of the standard part, including horizontal gradient projection and vertical gradient projection; S43, performing gradient traversal and variance filtering on the image of the standard part; S44, fitting the point set of the stable edge points of the image of the standard part; S45, calculating the actual size of the standard part in the image.

5. The camera stability monitoring method for battery vision inspection apparatus according to claim 4, wherein, The S42 specifically comprises: first determining the horizontal and vertical edges of the object in the image of the standard part, second calculating the gradient sum of each row for horizontal gradient projection, and then calculating the gradient sum of each column for vertical gradient projection, and using the following logic to perform gradient projection on the image of the standard part: where P horizontal is the total gradient intensity projection of the image in the horizontal direction, P vertical is the total gradient intensity projection of the image in the vertical direction, G x (x, y) is the gradient of position (x, y) in the input image in the x direction, G y (x, y) is the gradient of position (x, y) in the input image in the y direction.

6. The camera stability monitoring method for battery vision inspection apparatus according to claim 5, wherein, The S43 specifically comprises: first traversing the gradient value of the image to find the edge points whose gradient value exceeds a preset threshold, second calculating the variance in the neighborhood of the edge points, and if the variance is less than a set threshold, considering that the point is a stable edge point; if the variance is greater than the set threshold, considering that the point is an unstable edge point, removing the unstable edge points, and using the following logic to calculate the variance in the neighborhood of the edge points: where σ 2 is the variance, N is the number of points in the neighborhood, x i is the value of the ith point, and μ is the average value of the points in the neighborhood.

7. The camera stability monitoring method for battery vision inspection apparatus according to claim 6, wherein, The S44 specifically comprises: using least squares fitting on the stable edge points to fit the straight line of the edge of the standard part, and the straight line fitting mode is y=ax+b, wherein y represents, x represents, a represents the slope, and b represents the intercept, and using the following logic to solve the slope a and the intercept b: where x i and y i are the horizontal and vertical coordinate values of the i-th edge point.

8. The camera stability monitoring method for battery vision inspection apparatus according to claim 7, wherein, The S45 uses the following logic to calculate the actual size of the standard part in the image: Wherein, (x, y) is a point coordinate selected on the edge line of the standard part, N ′ pixels is the pixel number of the standard part in the image.

9. The camera stability monitoring method for battery vision inspection apparatus according to claim 1, wherein, The S5 uses the following logic to calculate the camera stability Stability Change: Stability Change = (N' pixels *P pixel )-D actual where P pixel is the pixel precision, D actual is the actual size, N′ pixels is the number of pixels of the standard part in the image, and if the Stability Change is closer to 0, it means that the camera stability is better.

10. A camera stability monitoring system for a battery vision inspection apparatus, characterized by, It comprises: A hardware device module for installing a standard part in the field of view of a camera, and capturing an image of the standard part by the camera; An image calibration module for calibrating the image of the standard part based on a pixel accuracy algorithm; A matching positioning module for accurately positioning the position of the standard part in the image based on a template matching algorithm; A stability monitoring module for measuring the size of the standard part in the image based on an edge gradient size detection algorithm, and monitoring the stability of the camera; If the detected size of the standard part in the image remains constant, it indicates that the camera is stable; If the detected size of the standard part in the image fluctuates, it indicates that the camera stability is affected; a stability evaluation module for comparing the measured actual size with the actual size of the known standard part, any deviation in the standard part size being taken as a change in the camera stability, and the camera stability being continuously monitored.

Images