Battery pack frame appearance quality inspection method and system based on machine vision

Abstract

The invention discloses a battery pack frame appearance quality inspection method and system based on machine vision, relates to the technical field of battery pack detection, and solves the problem that a traditional quality inspection method cannot meet the requirements of industrial development. According to the central point and the travel path direction, analyzing a pixel value sorting sequence and a ratio column, and accurately identifying an abnormal section and an abnormal area; according to the method, the limitation of traditional single point position comparison is abandoned, from the perspective of overall feature change, the accuracy of abnormal point position determination is improved, abnormal area distribution can be comprehensively displayed, visual and detailed abnormal information is provided for external personnel, rapid positioning of problem roots is assisted, quality inspection efficiency and problem solving speed are greatly improved, and the method is suitable for popularization and application. Comprehensiveness, accuracy and high efficiency of appearance quality detection of the battery pack frame are effectively guaranteed, and the method is of great significance to improvement of the overall quality of the battery pack.

Description

Technical Field

[0001] The present invention relates to the technical field of battery pack detection, and specifically to a method and system for quality inspection of the outer shape of a battery pack frame based on machine vision. Background Art

[0002] In the fields of new energy vehicles and energy storage systems, etc., the battery pack, as the core energy storage component, its performance and safety are crucial; and the battery pack frame, as the key structure for protecting the battery module and carrying components such as electrical connection parts and cooling systems, its outer shape quality directly affects the overall performance and reliability of the battery pack.

[0003] Traditional methods for quality inspection of the outer shape of the battery pack frame mainly rely on manual visual inspection and simple measuring tools; manual visual inspection is limited by the inspector's experience, fatigue level, and subjective judgment, making it difficult to ensure the consistency and accuracy of the detection results, and it is easy to miss some minor defects, such as fine scratches, cracks, etc.; moreover, for the outer shape of frames with complex shapes and large areas, manual inspection efficiency is extremely low and cannot meet the requirements of large-scale production.

[0004] Although simple measuring tools can obtain some dimensional information, they can only measure specific key dimensions and are difficult to comprehensively and real-time reflect the overall quality status of the frame outer shape; for some subtle changes in surface flatness, edge contours, and potential defects in the internal structure, traditional measuring tool measurement methods are even more powerless.

[0005] With the rapid development of the new energy industry, the production volume of battery packs has increased significantly, and the requirements for the outer shape quality of battery pack frames have also been increasing day by day; traditional quality inspection methods can no longer meet the needs of industrial development, and there is an urgent need for an efficient, accurate, and comprehensive quality inspection method to ensure that the outer shape of the battery pack frame meets the design standards and improve the overall quality and safety of the battery pack. Therefore, a method for quality inspection of the outer shape of a battery pack frame based on machine vision has emerged, which can overcome the limitations of traditional methods and provide strong support for the development of the battery pack industry. Summary of the Invention

[0006] In view of the deficiencies of the prior art, the present invention provides a method and system for quality inspection of the outer shape of a battery pack frame based on machine vision, which solves the problem that traditional quality inspection methods can no longer meet the needs of industrial development.

[0007] To achieve the above object, the present invention is realized through the following technical solutions: A method for quality inspection of the outer shape of a battery pack frame based on machine vision, including the following steps:

[0008] Step 1: Obtain the frame contour image of the battery pack based on a machine vision device, perform edge processing on the obtained frame contour image, use the Sobel algorithm to confirm the edge contour, and divide the frame contour image into multiple verification images. The specific sub-steps are as follows:

[0009] Perform grayscale processing on the obtained frame contour image to confirm the grayscale image associated with this image;

[0010] Calibrate the pixel values associated with different pixel points in the grayscale image as X i , where i represents different pixel points, and then use the Sobel algorithm to confirm the comprehensive gradient associated with the corresponding pixel points. Pixel points that satisfy: comprehensive gradient > Y1 are calibrated as gradient pixel points, where Y1 is a preset value. If not satisfied, no calibration is performed;

[0011] Connect several consecutive gradient pixel points to confirm the gradient contour, calibrate the different regions included in the gradient contour as different region contours, and confirm the image area of the different region contours. Region contours that satisfy: image area ∈ preset interval are denoted as characteristic contours, and the partial image associated with the characteristic contour is calibrated as a verification image, where the endpoint values of the preset interval are both preset values;

[0012] Step 2: Perform feature verification on the multiple verified verification images. Confirm the center point of this image from the verification image, and then confirm the deviation pixel points outward based on this center point, so as to lock the verification feature quantity line belonging to this verification image. Based on this verification feature quantity line, evaluate whether there is an error in this verification image and calibrate it as an error image. The specific method is as follows:

[0013] P1. Based on the edge contour of this verification image and combined with a two-dimensional coordinate system, confirm the center point of this verification image;

[0014] P2. Denote the pixel value associated with the center point as the initial pixel value, and denote the pixel values associated with other pixel points within the surrounding circle of this center point as other pixel values. Use: |initial pixel value - other pixel value| = pixel difference to confirm the different pixel differences associated with different other pixel points, select the maximum value from the different pixel differences, and use the other pixel point associated with the maximum value as the selected point. If there are multiple groups of selected points, randomly confirm one group of selected points from the multiple groups of selected points;

[0015] P3. Then use the surrounding circle as the inner circle, confirm its adjacent surrounding outer circle, perform difference processing on the pixel values between other pixel points within the surrounding outer circle and the selected point, confirm the corresponding pixel differences, and select the maximum value from the multiple groups of confirmed pixel differences to determine the selected point. If there are multiple groups of selected points, confirm the point closest to the selected point in the inner circle from the multiple groups of selected points as the determined selected point;

[0016] P4. Then, take the outer surrounding as the inner circle and re-confirm the outer surrounding. Repeat P3 to sequentially confirm different selected points associated within the subsequent outer surrounding until the determined selected point is located on the edge contour of the calibration image, and then stop.

[0017] P5. Starting from the center point, connect the sequentially determined selected points to confirm the calibration feature quantity line belonging to this calibration image.

[0018] P6. For the preset standard image, adopt the same processing method as P1 - P5. Confirm the calibration feature quantity line from the standard image and denote it as the standard line. Compare the calibration feature quantity line with the standard line to make the starting points of the two groups of lines coincide. Then, by rotating the calibration feature quantity line, identify whether there is a coincidence between the calibration feature quantity line and the standard line during the rotation process. If not, denote this calibration image as an error image. If there is a coincidence, no processing is required.

[0019] Step 3. Perform calibration analysis on the calibrated error image and the standard image to confirm the change characteristics in the same area of the error image and the standard image. Based on the feature differences between the change characteristics, lock the abnormal segments, denote the area covered by several abnormal segments as the abnormal area, and display the confirmed abnormal area. The specific method is as follows:

[0020] Calibrate the error image and the standard image to make the error image coincide with the standard image.

[0021] Take the center points calibrated in the error image and the calibrated image as the starting points, and randomly determine a set of directions as the traveling directions. Denote the other pixel points associated with the traveling directions as traveling pixel points. Starting from the starting points, sort the sequentially associated traveling pixel points and sort the pixel values associated with the corresponding positions to confirm the pixel value sorting sequence. Denote the pixel value sorting sequence associated with the error image as the error sequence, and denote the pixel value sorting sequence associated with the calibrated image as the calibration sequence.

[0022] Determine the absolute value of the pixel difference between adjacent pixel values in the error sequence, and according to the sorting method of the error sequence, sort the determined absolute values. Then, according to the sorting method, perform ratio processing on several groups of absolute values to confirm a set of error ratio columns. Process the calibration sequence in the same way to confirm a set of calibration ratio columns.

[0023] Identify whether the error ratio column and the calibration ratio column are exactly the same. If not, adjust the first ratios of the error sequence and the calibration sequence to the same ratio, and compare and verify the adjusted error sequence and calibration sequence. If they are exactly the same, no processing is required, and process the error sequences and calibration sequences associated with other traveling directions:

[0024] Identify the different ratios associated with the same sorting positions within, denote such ratios as error ratios, denote the pixel differences associated with the error ratios as error differences, denote the two sets of pixel points associated with this error difference as error points, and denote the value segment associated with the two sets of error points as an abnormal segment;

[0025] Determine the abnormal segments associated with several traveling directions in the error image in sequence, denote the area covered by adjacent abnormal segments as an abnormal area, and mark the abnormal area in the error image.

[0026] Preferably, a quality inspection system for the outer shape of a battery pack frame based on machine vision includes:

[0027] A machine vision terminal, which acquires the outer shape image of the battery pack frame, performs edge processing on the acquired outer shape image, uses the Sobel algorithm to confirm the edge contour, and divides the outer shape image into multiple calibration images;

[0028] An error image calibration terminal, which performs feature calibration on the multiple confirmed calibration images, determines the center point of this image from the calibration image, and then confirms the deviation pixel points outward from this center point, thereby locking the calibration feature quantity line belonging to this calibration image, and evaluates whether there is an error in this calibration image based on this calibration feature quantity line, and calibrates it as an error image;

[0029] An abnormal area display terminal, which performs calibration analysis on the calibrated error image and the standard image, confirms the change characteristics in the same area of the error image and the standard image, locks the abnormal segment based on the feature differences between the change characteristics, denotes the area covered by several abnormal segments as an abnormal area, and displays the confirmed abnormal area.

[0030] The present invention provides a quality inspection method and system for the outer shape of a battery pack frame based on machine vision. Compared with the prior art, it has the following beneficial effects:

[0031] The present invention determines the center point of the calibration image, and gradually screens the selected points according to the pixel value difference, and finally constructs the calibration feature quantity line. This method can sensitively capture the internal point distribution and numerical feature changes of the image; compared with the calibration feature quantity line of the standard image, it can accurately judge whether there is an error in the calibration image, quickly lock the error image, realize the efficient detection of the subtle differences in the local area of the outer shape of the battery pack frame, significantly improve the quality inspection accuracy, and avoid the error omission caused by manual subjective judgment;

[0032] After calibrating the error image with the standard image, analyzing the pixel value sorting sequence and ratio column based on the center point and the path direction can accurately identify abnormal segments and abnormal areas; this method abandons the limitations of traditional single-point comparison, starting from the perspective of overall feature changes, not only improves the accuracy of determining abnormal points, but also can comprehensively display the distribution of abnormal areas, providing intuitive and detailed abnormal information for external personnel, helping to quickly locate the root cause of problems, greatly improving the quality inspection efficiency and the speed of problem solving, effectively ensuring the comprehensiveness, accuracy and efficiency of the battery pack frame shape quality inspection, and having important significance for improving the overall quality of the battery pack. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic flowchart of the method of the present invention;

[0034] Figure 2 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] The First Embodiment

[0037] Please refer to Figure 1 , this application provides a method for inspecting the shape of the battery pack frame based on machine vision, including the following steps:

[0038] Step 1: Obtain the frame shape image of the battery pack based on the machine vision device, perform edge processing on the obtained frame shape image, use the Sobel algorithm to confirm the edge contour, and divide the frame shape image into multiple calibration images. Specifically, the Sobel algorithm can effectively detect the contour edges existing in the corresponding image. By confirming the gradient pixel points associated with the edges, the gradient edge contours associated with the corresponding partitions are confirmed, and then based on the different partitions associated with the different gradient edge contours, multiple calibration images are locked. There are multiple rectangular partition areas with the same area on the outer end face of the battery pack. After each partition area is calibrated, it is divided into the corresponding calibration image, which is convenient for subsequent overall quality inspection of the shape;

[0039] Among them, the specific sub-steps for dividing the calibration image are:

[0040] The acquired frame contour image is grayscale processed to confirm the grayscale image associated with this image (each point in its image is associated with a corresponding RGB value, different parameters in its RGB value are assigned different weight factors, and then based on the weight factors and the associated specific values, the grayscale value associated with the corresponding point is confirmed, and then based on the grayscale values associated with different points, the grayscale image is reconstructed to determine the corresponding grayscale image. Such a processing method is relatively common in the prior art, so it will not be elaborated here);

[0041] Calibrate the pixel values associated with different pixel points in the grayscale image as X i , where i represents different pixel points, and then use the Sobel algorithm to confirm the comprehensive gradient associated with the corresponding pixel point (confirmed by combining the vertical gradient and the vertical gradient). Pixel points that satisfy: comprehensive gradient > Y1 are calibrated as gradient pixel points, where Y1 is a preset value, and its specific value is determined by the operator according to experience. Pixel points that do not meet the corresponding conditions are not calibrated;

[0042] Connect several consecutive gradient pixel points (that is, gradient pixel points that appear in sequence and are adjacent to each other in turn), confirm the gradient contour (the gradient contour is an enclosing contour, and the area contained within the gradient contour is the corresponding contour area), and calibrate the different areas contained within the gradient contour as different area contours, and confirm the image area of the different area contours. Areas that satisfy: image area ∈ preset interval are recorded as characteristic contours, and the partial image associated with the characteristic contour is calibrated as the verification image. The endpoint values of the preset interval are both preset values, and the preset interval is determined in advance by the operator according to the area size of the actual verification area, that is, multiple different rectangular areas existing on the surface of the battery pack;

[0043] Step 2: Perform characteristic verification on the confirmed multiple verification images. Confirm the center point of this image from within the verification image, and then confirm the deviation pixel points outward from this center point, so as to lock the verification characteristic quantity line belonging to this verification image. Based on this verification characteristic quantity line, evaluate whether there is an error in this verification image and calibrate it as an error image. The specific calibration method is as follows:

[0044] P1. Based on the edge contour of this verification image, combined with the two-dimensional coordinate system, confirm the center point of this verification image (combine the edge contour with the two-dimensional coordinate system to confirm the different two-dimensional coordinates associated with different contour points, then perform mean processing on several groups of two-dimensional coordinates to confirm the mean coordinate, and based on the position of the mean coordinate, calibrate it within the verification image to confirm the center point of this verification image);

[0045] P2. Denote the pixel value associated with the center point as the initial pixel value, and denote the pixel values associated with other pixel points within the surrounding circle around this center point as other pixel values. Use: |initial pixel value - other pixel value| = pixel difference to confirm the different pixel differences associated with different pixel points. Select the maximum value from the different pixel differences, and use the other pixel point associated with the maximum value as the selected point. If there are multiple groups of selected points, randomly confirm one group of selected points from the multiple groups of selected points;

[0046] P3. Then use the surrounding circle as the inner circle, confirm its adjacent surrounding outer circle, and perform difference processing on the pixel values between other pixel points within the surrounding outer circle and the selected point to confirm the corresponding pixel difference, and select the maximum value from the multiple groups of confirmed pixel differences to determine the selected point. If there are multiple groups of selected points, confirm the point closest to the selected point in the inner circle from the multiple groups of selected points as the determined selected point;

[0047] P4. Then use the surrounding outer circle as the inner circle, and re-confirm the surrounding outer circle, repeat P3, and sequentially confirm the different selected points associated with the subsequent surrounding outer circles until the determined selected point is located on the edge contour of the calibration image;

[0048] P5. Starting from the center point, connect the sequentially determined selected points (that is, starting from the center point and gradually connecting the determined selected points outward to confirm a group of connecting lines) to confirm the calibration feature quantity line belonging to this calibration image;

[0049] P6. Use the same processing method as P1 - P5 for the preset standard image, confirm the calibration feature quantity line from the standard image and denote it as the standard line. Compare the calibration feature quantity line with the standard line to make the starting points of the two groups of lines coincide. Then, by rotating the calibration feature quantity line (rotating according to the position of the starting point), identify whether there is a coincidence between the calibration feature quantity line and the standard line during the rotation process. If there is, no processing is required. If not, denote this calibration image as an error image;

[0050] Specifically, when there is an error image, it means that the arrangement method and numerical characteristics between its internal points are different from those of the standard image. When there is such a difference, it will cause a large change in the difference between the corresponding pixel points. In the case of a large change difference, error confirmation is required to lock the corresponding error area and perform subsequent calibration. Based on the change situation of the pixel points within the corresponding image and the change situation of the standard image, it is possible to quickly determine whether there is a large error in this calibration image and make a comprehensive determination to complete the calibration process of the error image;

[0051] Step 3: Perform calibration analysis on the calibrated error image and the standard image, confirm the change characteristics in the same area within the error image and the standard image, lock the abnormal segments based on the feature differences between the change characteristics, and record the areas covered by several abnormal segments as abnormal areas, and display the confirmed abnormal areas for external personnel to view. The specific method for confirming the abnormal area is as follows:

[0052] Calibrate the error image and the standard image to make the error image coincide with the standard image (according to the image pose associated in the image acquisition process and the associated direction calibrated in the standard image, when coinciding, perform coincidence verification on the angles at the corresponding image edges to make them consistent);

[0053] Use the calibrated center points in the error image and the calibrated image as the starting points, and randomly determine a set of directions as the traveling directions. Denote the other pixel points associated with the traveling directions as traveling pixel points. Sort the sequentially associated traveling pixel points starting from the starting point, and sort the pixel values associated with the corresponding positions. Confirm the pixel value sorting sequence. Denote the pixel value sorting sequence associated in the error image as the error sequence, and denote the pixel value sorting sequence associated in the calibrated image as the calibration sequence (the traveling directions associated with the error sequence and the calibration sequence are exactly the same). For example, there are calibrated center points in the corresponding error image and the calibrated image. It is stipulated that the due north direction is used as the traveling direction. Starting from the center point, sort several pixel points located in the same set of traveling directions to confirm several pixel points in the same traveling direction. Each different pixel point is associated with a different pixel value, and the pixel values between adjacent pixel points can be used to confirm the difference, lock the corresponding pixel difference, and then perform ratio processing on several pixel differences to lock the corresponding ratio sorting sequence. Subsequently, in the comparison and calibration process with the calibrated image, confirm the different ratios existing inside it and lock the error ratio;

[0054] Determine the absolute value of the pixel difference associated with adjacent pixel values in the error sequence, sort the determined absolute values according to the sorting method of the error sequence, and perform ratio processing on several groups of absolute values according to the sorting method to confirm a column of error ratios. Process the calibration sequence in the same way to confirm a column of calibration ratios;

[0055] Identify whether the error ratio column is exactly the same as the calibration ratio column. If they are the same, no processing is required, and process the error sequence and the calibration sequence associated with other traveling directions. If they are not the same, adjust the first ratios of the error sequence and the calibration sequence to the same ratio (after adjustment, the other ratios associated with the corresponding ratio column will also change synchronously). Compare and calibrate the adjusted error sequence and calibration sequence:

[0056] Confirm the different ratios associated with the same sorting positions within, denote such ratios as error ratios, denote the pixel differences associated with the error ratios as error differences, denote the two sets of pixel points associated with this error difference as error points, and denote the value segment associated with the two sets of error points as an abnormal segment;

[0057] Successively determine the abnormal segments associated with several traveling directions within the error image, denote the area covered by adjacent abnormal segments as an abnormal area, calibrate the abnormal area within the error image, and display the calibrated error image for external personnel to view;

[0058] According to this determination process, the abnormal points with abnormalities in the corresponding traveling direction can be determined. According to the way of the change of the difference characteristics, the points with characteristic changes can be effectively confirmed. Compared with the original single-point comparison method, its accuracy is higher and a better determination effect can be achieved.

[0059] Second Embodiment

[0060] Combined with Figure 2 , a battery pack frame outer shape quality inspection system based on machine vision, including:

[0061] A machine vision end, which acquires the frame outer shape image of the battery pack, performs edge processing on the acquired frame outer shape image, uses the Sobel algorithm to confirm the edge contour, and divides the frame outer shape image into multiple calibration images;

[0062] An error image calibration end, which performs feature calibration on the multiple confirmed calibration images, confirms the center point of this image from within the calibration image, and then confirms the deviation pixel points outward from this center point, thereby locking the calibration feature quantity line belonging to this calibration image, and evaluating whether there is an error in this calibration image based on this calibration feature quantity line and calibrating it as an error image;

[0063] An abnormal area display end, which performs calibration analysis on the calibrated error image and the standard image, confirms the change characteristics of the same area within the error image and the standard image, locks the abnormal segment based on the feature differences between the change characteristics, denotes the area covered by several abnormal segments as an abnormal area, and displays the confirmed abnormal area.

[0064] Some of the data in the above formula are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0065] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for inspecting the appearance quality of a battery pack frame based on machine vision, characterized in that, Including the following steps: Step 1: Obtain the frame contour image of the battery pack based on a machine vision device, perform edge processing on the obtained frame contour image, use the Sobel algorithm to confirm the edge contour, and divide the frame contour image into multiple verification images; Step 2: Perform feature verification on the multiple verified verification images, confirm the center point of this image from within the verification image, and then confirm the deviation pixel points outward from this center point, so as to lock the verification feature quantity line belonging to this verification image. Based on this verification feature quantity line, evaluate whether there is an error in this verification image and mark it as an error image; Step 3: Perform verification analysis on the calibrated error image and the standard image, confirm the change characteristics in the same area of the error image and the standard image, and based on the feature differences between the change characteristics, lock the abnormal segments, and record the area covered by several abnormal segments as the abnormal area, and display the confirmed abnormal area.

2. The method for inspecting the appearance of the battery pack frame based on machine vision according to claim 1, wherein In the above Step 1, the specific sub-steps for dividing the verification images are: Perform grayscale processing on the obtained frame contour image to confirm the grayscale image associated with this image; Calibrate the pixel values associated with different pixel points in the grayscale image as X i , where i represents different pixel points, and then use the Sobel algorithm to confirm the comprehensive gradient associated with the corresponding pixel points. The pixel points that satisfy: comprehensive gradient > Y1 are calibrated as gradient pixel points, where Y1 is a preset value; Connect several consecutive gradient pixel points to confirm the gradient contour, and mark the different areas included in the gradient contour as different area contours, and confirm the image areas of the different area contours. Mark the area contour that satisfies: the image area ∈ the preset interval as the feature contour, and mark the partial image associated with the feature contour as the verification image, and the end point values of the preset interval are both preset values.

3. The method for inspecting the outer shape of a battery pack frame based on machine vision according to claim 2, wherein Pixels that do not satisfy the comprehensive gradient > Y1 are not marked in any way.

4. The method for inspecting the outer shape quality of a battery pack frame based on machine vision according to claim 1, wherein, In the above Step 2, the specific method for calibrating the error image is: P1. Based on the edge contour of this verification image and combined with the two-dimensional coordinate system, confirm the center point of this verification image; P2. Denote the pixel value associated with the center point as the initial pixel value, and denote the pixel values associated with other pixel points within the surrounding circle of this center point as other pixel values. Use: |initial pixel value - other pixel value| = pixel difference to confirm the different pixel differences associated with different other pixel points, select the maximum value from the different pixel differences, and use the other pixel point associated with the maximum value as the selected point. If there are multiple groups of selected points, randomly confirm a group of selected points from the multiple groups of selected points; P3. Then use the surrounding circle as the inner circle, confirm its adjacent surrounding outer circle, and perform difference processing on the pixel values between other pixel points within the surrounding outer circle and the selected point to confirm the corresponding pixel difference, and select the maximum value from the confirmed multiple groups of pixel differences to determine the selected point. If there are multiple groups of selected points, confirm the point closest to the selected point in the inner circle from the multiple groups of selected points as the determined selected point; P4. Then use the surrounding outer circle as the inner circle, and re-confirm the surrounding outer circle, repeat P3, and sequentially confirm the different selected points associated with the subsequent surrounding outer circles until the determined selected point is located on the edge contour of the verification image and then stop; P5. Starting from the center point, connect the sequentially determined selected points to confirm the verification feature quantity line belonging to this verification image; For the preset standard image, adopt the same processing method as P1 - P5. Confirm the calibration feature quantity line within the standard image and denote it as the standard line. Compare the calibration feature quantity line with the standard line to make the starting points of the two groups of lines coincide. Then, by rotating the calibration feature quantity line, identify whether there is a coincidence between the calibration feature quantity line and the standard line during the rotation process. If not, denote this calibration image as an error image.

5. The method for inspecting the outer shape of a battery pack frame based on machine vision according to claim 4, wherein In step P6 described above, if there is a coincidence between the calibration feature quantity line and the standard line during the rotation process, no processing is required.

6. The method for inspecting the appearance of a battery pack frame based on machine vision according to claim 1, characterized in that, In step three described above, the specific method for confirming the abnormal area is as follows: Calibrate the error image and the standard image to make the error image coincide with the standard image; Using the center points marked in the error image and the calibrated image as the starting points, and randomly determining a group of directions as the traveling directions. Denote the other pixel points associated with the traveling directions as traveling pixel points. Starting from the starting points, sort the sequentially associated traveling pixel points, and sort the pixel values associated with the corresponding positions. Confirm the pixel value sorting sequence. Denote the pixel value sorting sequence associated with the error image as the error sequence, and denote the pixel value sorting sequence associated with the calibrated image as the calibration sequence; Determine the absolute value of the pixel difference associated with adjacent pixel values in the error sequence, and according to the sorting method of the error sequence, sort the determined absolute values. Then, according to the sorting method, perform a ratio process on several groups of absolute values to confirm a group of error ratio columns. Process the calibration sequence in the same way to confirm a group of calibration ratio columns; Identify whether the error ratio column is exactly the same as the calibration ratio column. If not, adjust the first ratios of the error sequence and the calibration sequence to the same ratio, and compare and verify the adjusted error sequence and calibration sequence: Confirm the different ratios associated with the same sorting positions inside, denote such ratios as error ratios, denote the pixel difference associated with the error ratio as the error difference, denote the two groups of pixel points associated with this error difference as error points, and denote the value segment associated with the two error points as the abnormal segment; Determine the abnormal segments associated with several traveling directions in the error image in sequence, denote the area covered by adjacent abnormal segments as the abnormal area, and mark the abnormal area in the error image.

7. The method for inspecting the outer shape quality of a battery pack frame based on machine vision according to claim 6, characterized in that, If the error ratio column is exactly the same as the calibration ratio column, no processing is required, and process the error sequences and calibration sequences associated with other traveling directions.

8. A battery pack frame outer shape quality inspection system based on machine vision, which operates according to the method for inspecting the outer shape of a battery pack frame based on machine vision described in any one of claims 1-7, characterized in that, Including: The machine vision terminal acquires the frame shape image of the battery pack, performs edge processing on the acquired frame shape image, uses the Sobel algorithm to confirm the edge contour, and divides the frame shape image into multiple calibration images; The error image calibration terminal performs feature calibration on the confirmed multiple calibration images, confirms the center point of this image from the calibration image, and then confirms the deviation pixel points outward based on this center point, thereby locking the calibration feature quantity line belonging to this calibration image. Based on this calibration feature quantity line, evaluate whether there is an error in this calibration image and mark it as an error image; The abnormal area display terminal performs calibration analysis on the calibrated error image and the standard image, confirms the change characteristics of the same area in the error image and the standard image, locks the abnormal segments based on the feature differences between the change characteristics, records the area covered by several abnormal segments as the abnormal area, and displays the confirmed abnormal area.

Images