Method and system for detecting the position of the tabs in a battery pole group based on machine vision

Abstract

The present invention relates to the field of battery technology, and in particular to a method and system for detecting the position of a tab in a battery pole group based on machine vision; the method comprises: obtaining a connected domain of each tab in a battery pole group image to be detected; extracting three features of the connected domain of each tab; calculating the similarity between any feature of each tab and the corresponding feature of each template tab in a template battery pole group image; using a KM algorithm to match each tab in the battery pole group image to be detected with each template tab in a template battery pole group image, obtaining a matching result and a sum of edge weights of the matching result; inputting the sum of edge weights into a first network model and a second network model respectively, obtaining a first abnormality rate and a second abnormality rate, and obtaining an abnormal probability; in response to the abnormal probability being greater than a threshold value, there is a tab with a position misalignment or defect in the battery pole group image to be detected. The present invention can accurately determine whether the tab in the battery pole group is qualified, thereby improving the reliability of the battery.

Description

Technical Field

[0001] The present invention relates to the field of battery technology, and more specifically, to a method and system for detecting the position of a tab in a battery pole group based on machine vision. Background Art

[0002] The battery pole group is the core component of the battery, which consists of positive and negative plates and separators. The positive plates and negative plates are arranged alternately, and the separator is set between the positive plates and negative plates to prevent them from directly contacting each other and causing short circuits. Among them, the battery tabs that connect the battery poles (the battery poles are one of the positive and negative poles of the battery) and the conductive elements of the battery assembly are the necessary medium for the transmission of current inside the battery to ensure the normal operation of the battery.

[0003] Specifically, the tab is a metal conductor that leads the positive and negative electrodes from the battery cell, that is, the contact point between the positive and negative electrodes of the battery during charging and discharging, and is an important component of the lithium-ion polymer battery. In the industrial manufacturing process, due to the influence of factors such as material quality, diaphragm thickness and cutting process, the tab may be misaligned or defective, which greatly affects the battery performance and even causes the battery quality to be unqualified. Therefore, how to detect whether the tab is misaligned is a problem that needs to be solved.

[0004] In the related technology, such as the Chinese patent application document with application publication number CN112184816A and titled A flexible positioning method for lithium batteries based on battery tabs, it is disclosed that the tab area of ​​the battery body is extracted, the tab area of ​​the battery body is screened, and the difference area is obtained. Each connected domain in the difference area is processed separately to obtain the tab area, the tab is positioned, and the position of the tab is obtained.

[0005] Although the above scheme can determine the position of the tab, each tab needs to be processed when determining the position of the tab. When faced with the position detection of a large number of battery tabs to determine whether they are qualified, there is obviously a problem of low efficiency. Summary of the invention

[0006] The purpose of the present invention is to propose a method and system for detecting the position of the tabs in a battery pole group based on machine vision, so as to solve the problem of low efficiency in the prior art when detecting the position of the tabs in a large number of battery pole groups. To this end, the present invention provides solutions in the following two aspects.

[0007] In a first aspect, the present invention provides a method for detecting the position of a tab in a battery pole group based on machine vision, comprising:

[0008] Obtaining the connected domain of each pole ear in the pole group image of the battery to be detected;

[0009] Extracting three features of the connected domain of each tab, wherein the three features are area, brightness feature sequence and position feature; calculating the similarity between any feature of each tab and the corresponding feature of each template tab in the template battery pole group image;

[0010] The KM algorithm is used to match each pole ear in the battery pole group image to be detected with each template pole ear in the template battery pole group image to obtain a matching result, and the sum of the edge weights of the matching result is obtained; wherein during matching, the edge weight between each pole ear and the template pole ear is the mean of the similarity of all corresponding features;

[0011] Inputting the edge weight sum into the first network model and the second network model respectively, to obtain a first abnormality rate and a second abnormality rate respectively;

[0012] The first abnormality rate and the second abnormality rate are integrated according to the accuracy of the historical matching results to obtain the abnormality probability; in response to the abnormality probability of the battery pole group image to be detected being greater than a threshold, there are misplaced or defective pole ears in the battery pole group to be detected.

[0013] The above scheme extracts multiple features of each tab in the battery pole group image to be detected, calculates the similarity between the tabs and the template tabs in the template battery pole group image, and provides edge weight data for subsequent KM matching, thereby improving the accuracy of the matching results; at the same time, starting from the matching results, different network models are used to estimate the overall abnormality rate of the tabs in the battery pole group image to be detected, and finally, combined with prior information, different abnormality rates are integrated to obtain the final abnormality probability, so as to judge whether there are any unqualified tabs in the current battery pole group to be detected as a whole, thereby improving the detection efficiency.

[0014] Optionally, the brightness feature sequence is composed of the maximum brightness values ​​corresponding to each row or column in the connected domain corresponding to each pole lug; the position feature is calculated by calculating the difference between the maximum distance and the minimum distance between the position of any pole lug and the positions of multiple adjacent pole lugs, and the position is the position when the brightness value in the corresponding pole lug is the largest; the area is the number of pixel points in the connected domain corresponding to the pole lug.

[0015] The above scheme obtains the area, brightness feature sequence and position features of each tab, which can characterize the properties of the tab from different angles and provide data support for subsequent matching.

[0016] Optionally, when the feature is a brightness feature sequence, the similarity is the Pearson correlation coefficient between the brightness feature sequence of any pole lug and the brightness feature sequence of any template pole lug in the template battery pole group image.

[0017] Optionally, the first network model training process is:

[0018] Obtain a first training data set, wherein the first training data set includes the sum of historical edge weights and labels when the matching results in the historical records are correct, wherein the labels are whether the tab position is normal or abnormal; train a first network model using the first training data set to obtain a trained first network model;

[0019] The training process of the second network model is:

[0020] Obtain a second training data set, wherein the second training data set includes the sum of historical edge weights and labels when the matching results in the historical records are wrong, wherein the label is whether the position of the tab is normal or abnormal; use the second training data set to train a second network model to obtain a trained second network model.

[0021] The two network models of the above scheme are trained using data with correct matching results and data with incorrect matching results respectively. The model training is performed from two opposite angles, considering the probability of position anomaly in the cases of correct matching and incorrect matching. The probabilities of position anomaly in different situations are integrated together according to the accuracy of historical matching results, which can improve the accuracy of identifying tab position anomaly.

[0022] Optionally, the first network model adopts a random forest model; the second network model adopts a random forest model or a support vector machine.

[0023] Optionally, the accuracy of the historical matching results is the ratio of the number of correct matching results to all matching results when the tabs in multiple historical battery pole groups are matched with the template tabs in the template battery pole group image respectively.

[0024] Optionally, the abnormal probability is:

[0025] ;

[0026] in, represents the abnormal probability of the battery pole group image to be detected, Indicates the accuracy of historical matching results. represents the first abnormality rate, Represents the second anomaly rate.

[0027] The above scheme uses the prior accuracy as a weight to evaluate the importance of anomaly rates under different models, thereby improving the accuracy of the final result.

[0028] Optionally, the process of obtaining the connected domain of each pole lug in the pole group image of the battery to be detected is:

[0029] Grayscale processing is performed on the image of the battery pole group to be inspected to obtain a grayscale image;

[0030] The battery pole group region in the grayscale image is extracted by adopting the threshold segmentation method;

[0031] The connected domain analysis of the battery pole group area is performed to obtain the connected domain of each pole ear.

[0032] The above solution can accurately obtain the area of ​​each pole ear in the image of the battery pole group to be inspected.

[0033] In a second aspect, the present invention provides a battery pole group lug position detection system based on machine vision, comprising:

[0034] processor;

[0035] A memory storing computer instructions for detecting the position of the tabs in a battery pole group based on machine vision. When the computer instructions are executed by the processor, the system executes the above-mentioned method for detecting the position of the tabs in a battery pole group based on machine vision.

[0036] The beneficial effects of the present invention are:

[0037] The solution of the present invention can judge the overall abnormal probability of all the tabs in the battery group, so as to determine whether the tabs in the battery group have position errors or defects, thereby improving the efficiency and accuracy of the judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0039] Figure 1 The flowchart of the method for detecting the position of the tabs in the battery pole group based on machine vision in this embodiment is schematically shown;

[0040] Figure 2 The structural block diagram of the battery pole group tab position detection system based on machine vision in this embodiment is schematically shown. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0042] The specific scenario targeted by the present invention is: Since the battery tab is a conductive element connecting the battery pole piece (battery positive or negative pole) and the battery assembly, it is a necessary medium for the transmission of current inside the battery. A well-designed tab can prevent the negative pole and the positive pole from directly contacting each other, thereby preventing the occurrence of a short circuit, which is an important measure to maintain battery safety. Therefore, by detecting the position or defects of multiple tabs in the battery pole group, it is possible to determine whether the battery pole group is qualified.

[0043] like Figure 1 As shown, the method for detecting the position of the tabs in the battery pole group based on machine vision in this embodiment includes the following steps:

[0044] Step S1, obtaining a battery pole group region in a battery pole group image to be detected, and determining a connected domain of each pole lug in the battery pole group region.

[0045] In this embodiment, a CCD camera or a CMOS camera is used to capture an image of a battery pole group, and after the image is captured, a median filter is used to perform denoising on the captured image to obtain an image of the battery pole group to be detected.

[0046] Since the battery group image to be detected contains a background area and a target area (battery group area), it is necessary to extract the battery group area through a threshold segmentation method. Specifically, the battery group image obtained in step 1 is grayed to obtain a grayscale image, and the grayscale image is threshold segmented, and the values ​​greater than the threshold are set to 1 and values ​​less than the threshold are set to 0 to obtain a binary image. The binary image and the grayscale image are multiplied to obtain the battery group area.

[0047] Of course, as another implementation, the battery pole group region can also be extracted by edge detection method. Specifically, the edge detection method can be Canny algorithm, Sobel operator, etc.

[0048] In this embodiment, the connected domain of each pole lug is obtained by performing a connected domain analysis on the battery pole group region.

[0049] Step S2, based on the connected domain of each pole lug, determine the feature set of each pole lug; based on the feature set of each pole lug, use the KM algorithm to match each pole lug in the battery pole group image to be detected with each template pole lug in the template battery pole group image to obtain a matching result, and obtain the sum of the edge weights of the matching result.

[0050] Since all the tabs in the battery group are identified and analyzed in this embodiment, when identifying the tab positions in the battery group, it is necessary to match the battery group image to be identified with the template battery group image to see whether the matching degree meets the requirements. Therefore, when matching, it is necessary to obtain the feature set of each tab, where the feature set includes area, brightness feature sequence and position feature.

[0051] Among them, the area of ​​each pole ear is obtained as follows:

[0052] Connected domains are extracted from the battery pole group area to obtain the connected domains corresponding to each pole lug. Each connected domain is labeled from left to right and from top to bottom according to the captured image, and the number of pixel values ​​in each connected domain is counted to obtain the area of ​​the corresponding pole lug. If a pole lug has defects, such as uneven polishing, the size characteristics of the pole lug will be different from those of other normal pole lugs.

[0053] The method for obtaining the brightness characteristic sequence of the tab is as follows: convert the battery pole group area into the HSV color space, obtain the brightness value of each row of the connected domain to which each tab belongs according to the brightness value of the brightness space (V channel) in the color space, select the maximum brightness value in each row, and form a characteristic sequence with the maximum brightness values ​​corresponding to all rows; of course, as other implementation methods, the brightness value of each column can also be determined, and the maximum brightness values ​​corresponding to all columns can be formed into a characteristic sequence.

[0054] The brightness feature sequence of the connected domain of each tab can not only reflect the brightness of the connected domain to which the tab belongs, but also reflect the change of the brightness value in the connected domain to which the tab belongs.

[0055] The method for obtaining the position features of the tabs is as follows: The purpose of calculating the position features of the tabs is to prevent the tabs of different battery pole groups from matching together during matching, which may lead to errors in the detection of the pole group positions. Therefore, the distance between the tabs is calculated as the position features of the tabs. When calculating the position features of the tabs, it is not necessary to calculate the distance between each tab and all other tabs. It is only necessary to calculate the distance between each tab and the other tabs around it. If the position of a tab deviates, the deviation will be reflected in the distance between the tab and the other tabs around it. The specific calculation process is as follows:

[0056] According to the brightness value of the tab, obtain the brightness value sequence of all brightness values ​​corresponding to the tab, determine the position of the maximum brightness value in the brightness value sequence, and this position is the position of the tab. The position corresponding to the maximum brightness value generally appears at the center point of the tab area. Therefore, the position of the maximum brightness value is used as the position of the tab, and the distance between each tab and other surrounding tabs is calculated. The specific calculation formula is as follows:

[0057] ;

[0058] in, represents the position characteristics of the i-th tab, represents the set of other tabs around the i-th tab, Indicates the position corresponding to the maximum brightness value in the i-th ear area, Indicates the position corresponding to the maximum brightness value in the surrounding j-th ear area, express Norm.

[0059] In the present embodiment, the reason for using the maximum value minus the minimum value of the distance between a pole ear and its adjacent pole ears as the position characteristic of the pole ear is that when the position of a pole ear deviates, it will be close to a certain pole ear adjacent to it, but at the same time it will also be away from another pole ear around it. If the sum of the distances between the pole ear and its adjacent pole ears is used, it may appear that regardless of whether the position of the pole ear deviates, the sum of the distances between the pole ear and multiple adjacent pole ears is the same.

[0060] In one embodiment, the detection of the tabs in the battery group is crucial to the safety of the battery. If the tabs are installed incorrectly or have defects, safety problems such as short circuits will occur. Therefore, the accurate position of the tabs is verified during the production process to ensure product quality. In the process of using images to detect the tabs in the battery group, because the main materials of the tabs are metal materials such as copper, aluminum or nickel, the degree of reflection is relatively large, therefore, the brightness of the tabs is used to determine whether the tabs are qualified.

[0061] In this embodiment, when matching, the edge weight of the KM algorithm is also obtained, wherein the edge weight is obtained by:

[0062] Calculate the similarity between the area of ​​any pole ear and the size feature of any template pole ear in the template battery pole group image; calculate the similarity between the brightness feature sequence of any pole ear and the brightness feature sequence of any template pole ear in the template battery pole group image; calculate the similarity between the position feature of any pole ear and the position feature of any template pole ear in the template battery pole group image; and use the average of the three similarities as the edge weight of KM matching. Among them, the similarity of size feature and position feature can be calculated based on Euclidean distance, and the larger the Euclidean distance, the smaller the similarity; taking the size feature as an example, it can be the square root of the difference between the size feature of the pole ear and the size feature of the template pole ear. It should be noted that the size feature in this embodiment is the area.

[0063] The similarity of the brightness feature sequence can be obtained by calculating the Pearson correlation coefficient; of course, it can also be obtained by calculating the cosine similarity.

[0064] Specifically, the calculation formula for the similarity mean is as follows:

[0065] ;

[0066] in, The mean similarity between the sth test tab and the tth template tab, represents the similarity of the size characteristics between the sth test tab and the tth template tab, represents the similarity of the brightness feature sequence between the sth test lug and the tth template lug, Represents the similarity of the position features between the sth lug to be tested and the tth template lug.

[0067] After determining the sum of the edge weights between each pole lug and each standard pole lug, the KM algorithm can be used to match each pole lug in the battery pole group image to be detected with each template pole lug in the template battery pole group image to obtain the matching results and the sum of the edge weights corresponding to the matching results.

[0068] Among them, the KM (Kuhn-Munkres) algorithm is an algorithm for solving the optimal matching problem in a weighted bipartite graph. It is suitable for finding a perfect match with weights from one group of nodes to another group of nodes in a bipartite graph. Therefore, after the matching is completed, each pole ear in the battery pole group image to be detected and each template pole ear in the template battery pole group image are one-to-one. Since the KM algorithm is an existing technology, it will not be described in detail here.

[0069] The above-mentioned middle template battery pole group image is a qualified battery pole group among the historical battery pole groups, that is, the position of the pole ears is correct and there are no defects.

[0070] Step S3, inputting the edge weight sum into the first network model and the second network model respectively, and obtaining the first abnormality rate and the second abnormality rate of the edge weight sum respectively.

[0071] In this embodiment, the first network model may adopt a random forest model or a support vector machine; the second network model may adopt a random forest model or a support vector machine.

[0072] Exemplarily, when the first network model is a random forest model, the training process of the random forest model is:

[0073] A first training data set is obtained, wherein the first training data set includes the sum of historical edge weights and labels when the matching results in the historical records are correct, wherein the label indicates whether the position of the tab is normal or abnormal, and a random forest model is trained using the first training data set to obtain a trained random forest model.

[0074] When the second network model is a support vector machine, its training process is:

[0075] A second training data set is obtained, wherein the second training data set includes the sum of historical edge weights and labels when the matching results in the historical records are wrong, wherein the label is whether the position of the tab is normal or abnormal, and a support vector machine is trained using the second training data set to obtain a trained support vector machine.

[0076] In this embodiment, after obtaining the trained first network model and the second network model, when detecting the image of the battery pole group to be detected in real time, the sum of the edge weights of the real-time best matching result is respectively input into the first network model (the network model corresponding to the matching result when it is correct) and the second network model (the network model corresponding to the matching result when it is wrong) to obtain the first abnormality rate and the second abnormality rate.

[0077] It should be noted that the "first" and "second" mentioned above are only used to distinguish the functions, and do not limit the meaning of the model itself. As another implementation, the first network model can also be trained using a training set with incorrect matching results, and the second network model can also be trained using a training set with correct matching results.

[0078] Step S4, according to the accuracy of the historical matching results, the first abnormality rate and the second abnormality rate are integrated to obtain the abnormality probability; in response to the abnormality probability of the battery pole group image to be detected being greater than the threshold, there are unqualified tabs in the battery pole group to be detected. The unqualified tabs refer to the wrong position of the tabs or the existence of defects in the tabs.

[0079] In this embodiment, the accuracy of the historical matching results is the ratio of the number of correct matching results to all matching results when matching the tabs in the historical battery pole group image with the template tabs in the template battery pole group image. This ratio is the accuracy.

[0080] The abnormal probability in this embodiment is:

[0081] ;

[0082] in, represents the abnormal probability of the battery pole group image to be detected, Indicates the accuracy of historical matching results. represents the first abnormality rate, Represents the second anomaly rate.

[0083] Among them, when the abnormal probability of the real-time detected battery pole group image to be detected is greater than the threshold, it is considered that the position of the pole ear in the battery pole group to be detected at this time is wrong or the pole ear has defects, and timely maintenance is required.

[0084] The solution of the present invention analyzes the features of the tabs in the battery pole group image to be detected and the tabs in the template battery pole group image, and matches the tabs in the battery pole group image to be detected with the tabs in the template battery pole group image by means of the KM algorithm. It can judge the overall abnormal probability of all the tabs in the battery pole group, thereby determining whether there are position errors or defects in the tabs in the battery pole group, thereby improving the efficiency and accuracy of the judgment.

[0085] The present invention also provides a system for detecting the position of the tabs in a battery group based on machine vision. Figure 2 As shown, the system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for detecting the position of the tabs in a battery pole group based on machine vision according to the present invention is implemented.

[0086] The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, whose configuration and functions are known in the art and thus will not be described in detail here.

[0087] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0088] In the description of this specification, “plurality” means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0089] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for detecting the position of the tabs in a battery pole group based on machine vision, characterized in that: include: The connected domain of each pole ear in the pole group image of the battery to be detected is obtained, and the process is: grayscale processing is performed on the pole group image of the battery to be detected to obtain a grayscale image; The battery pole group region in the grayscale image is extracted by adopting a threshold segmentation method; the connected domain of the battery pole group region is analyzed to obtain the connected domain of each pole lug; Extracting three features of the connected domain of each tab, wherein the three features are area, brightness feature sequence and position feature; calculating the similarity between any feature of each tab and the corresponding feature of each template tab in the template battery pole group image; The brightness feature sequence is composed of the maximum brightness values ​​corresponding to each row or column in the connected domain corresponding to each lug; The position feature is to calculate the difference between the maximum value and the minimum value of the distance between the position of any pole lug and the positions of multiple pole lugs adjacent to it, and the position is the position at which the brightness value in the corresponding pole lug is the largest; the area is the number of pixel points in the connected domain corresponding to the pole lug; The KM algorithm is used to match each pole ear in the battery pole group image to be detected with each template pole ear in the template battery pole group image to obtain a matching result, and the sum of the edge weights of the matching result is obtained; wherein during matching, the edge weight between each pole ear and the template pole ear is the mean of the similarity of all corresponding features; Inputting the edge weight sum into the first network model and the second network model respectively, to obtain a first abnormality rate and a second abnormality rate respectively; The first abnormality rate and the second abnormality rate are integrated according to the accuracy of the historical matching results to obtain the abnormality probability; in response to the abnormality probability of the battery pole group image to be detected being greater than a threshold, there are misplaced or defective pole ears in the battery pole group to be detected.

2. The method for detecting the position of the tabs in a battery pole group based on machine vision according to claim 1, characterized in that: When the feature is a brightness feature sequence, the similarity is the Pearson correlation coefficient between the brightness feature sequence of any pole lug and the brightness feature sequence of any template pole lug in the template battery pole group image.

3. The method for detecting the position of the tabs in a battery pole group based on machine vision according to claim 1, characterized in that: The first network model training process is: Obtain a first training data set, wherein the first training data set includes the sum of historical edge weights and labels when the matching results in the historical records are correct, wherein the labels are whether the tab position is normal or abnormal; train a first network model using the first training data set to obtain a trained first network model; The training process of the second network model is: Obtain a second training data set, wherein the second training data set includes the sum of historical edge weights and labels when the matching results in the historical records are wrong, wherein the label is whether the position of the tab is normal or abnormal; use the second training data set to train a second network model to obtain a trained second network model.

4. The method for detecting the position of the tabs in a battery pole group based on machine vision according to claim 3, characterized in that: The first network model adopts a random forest model; the second network model adopts a random forest model or a support vector machine.

5. The method for detecting the position of the tabs in a battery pole group based on machine vision according to claim 1, characterized in that: The accuracy of the historical matching results is the ratio of the number of correct matching results to all matching results when the tabs in multiple historical battery pole group images are matched with the template tabs in the template battery pole group image respectively.

6. The method for detecting the position of the tabs in a battery pole group based on machine vision according to claim 5, characterized in that: The abnormal probability is: ； in, represents the abnormal probability of the battery pole group image to be detected, Indicates the accuracy of historical matching results. represents the first abnormality rate, Represents the second anomaly rate.

7. The battery pole group lug position detection system based on machine vision is characterized by: include: processor; A memory storing computer instructions for detecting the position of a tab in a battery pole group based on machine vision, wherein when the computer instructions are executed by the processor, the system executes the method for detecting the position of a tab in a battery pole group based on machine vision according to any one of claims 1 to 6.

Images