Detection method and device, storage medium and program product

Abstract

The invention discloses a detection method and device, a storage medium and a program product. The detection device can obtain a cathode image and an anode image of a battery; performing target part identification processing on the cathode image and the anode image by using a trained target identification algorithm to obtain an identification result of the target part of the battery; wherein the trained target recognition algorithm is obtained by training an initial target recognition algorithm based on images of different batteries; and carrying out coordinate extraction processing on the identification result to obtain a position detection result of the target part.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a detection method, apparatus, storage medium, and program product. Background Technology

[0002] In the battery production process, the positioning of the cover plate and the adapter plate is involved. Currently, industrial cameras such as charge-coupled devices (CCDs) are mainly used on the production line to capture images, and then the adapter plate and the cover plate are detected and positioned by Hough transform and template matching. However, due to the variability of lighting, the diversity of materials, and the instability of processes in actual production, these factors can interfere with the above detection methods, thereby affecting the accuracy of the detection. Summary of the Invention

[0003] This application provides a detection method, apparatus, storage medium, and program product that can effectively improve the accuracy of detection.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] In a first aspect, embodiments of this application provide a detection method, the method comprising:

[0006] Acquire images of the battery's cathode and anode;

[0007] The trained target recognition algorithm is used to identify the target parts of the cathode image and the anode image to obtain the identification results of the target parts of the battery; wherein, the trained target recognition algorithm is obtained by training the initial target recognition algorithm based on images of different batteries;

[0008] The recognition results are processed by coordinate extraction to obtain the position detection results of the target part.

[0009] In this embodiment, the detection device can train a target recognition algorithm using images of different batteries, and then use the trained target recognition algorithm to perform target recognition on the cathode and anode images of the battery. This can be widely applied to target recognition of different types of batteries, effectively improving the anti-interference and robustness of target part detection on the battery. After obtaining the recognition result of the target part, the coordinates of the recognition result can be directly extracted to determine the position detection result of the target part to be detected. This enables precise positioning of the target part in the battery, greatly improving the accuracy of detection.

[0010] In some embodiments of this application, the target location includes at least one of an adapter plate and a cover plate; the step of performing coordinate extraction processing on the identification result to obtain the position detection result of the target location includes:

[0011] Based on the recognition result of the adapter piece, the coordinates of the center point of the target area are extracted to obtain the position detection result of the adapter piece;

[0012] Based on the recognition result of the cover plate, the coordinates of the cover plate edge line are extracted to obtain the position detection result of the cover plate.

[0013] In this embodiment, the adapter plates and cover plates of the battery cathode and anode can be detected and located. This includes extracting the coordinates of the center point of the target area after obtaining the recognition result of the adapter plate using a trained target recognition algorithm to determine the position detection result of the adapter plate, and extracting the coordinates of the edge line of the cover plate after obtaining the recognition result of the cover plate using a trained target recognition algorithm to determine the position detection result of the cover plate. This achieves accurate detection and location of the adapter plates and cover plates in the battery.

[0014] In some embodiments of this application, the target area is a circular area corresponding to the pads in the adapter piece, and the position detection result of the adapter piece includes the position detection result of the center point of the pads; the step of extracting the coordinates of the center point of the target area based on the identification result of the adapter piece to obtain the position detection result of the adapter piece includes:

[0015] The circular region is fitted based on the first contour point in the recognition result of the adapter piece to obtain the fitted circular region.

[0016] Determine the center coordinates of the fitted circular region, and use the center coordinates as the position detection result of the center point of the pad.

[0017] In this embodiment, when determining the position detection result of the adapter piece, the main focus is on detecting the circular area corresponding to the pad in the adapter piece. During the detection process, the circular area is fitted, and then the center point position of the pad is determined based on the center coordinates of the fitted circular area, thereby achieving effective detection of the position of the adapter piece in the battery.

[0018] In some embodiments of this application, the step of extracting the coordinates of the cover plate edge line based on the recognition result of the cover plate to obtain the position detection result of the cover plate includes:

[0019] The ordinate values ​​of the second contour points in the identification results of the cover plate are sorted in descending order to obtain the sorting result;

[0020] Based on the sorting result, a first number of third contour points and a first number of fourth contour points are determined; wherein, the minimum ordinate value of the third contour point is greater than the maximum ordinate value of the fourth contour point;

[0021] The position detection result of the cover plate is obtained based on the third contour point and the fourth contour point.

[0022] In this embodiment, when determining the position detection result of the cover plate, the ordinate values ​​of the second contour points in the recognition result of the cover plate can be sorted in descending order, that is, from largest to smallest. Then, based on the sorting result, the first number of third contour points and the first number of fourth contour points are selected. Finally, the position detection result of the cover plate is determined by using the third contour points and the fourth contour points, which can accurately locate the position of the cover plate.

[0023] In some embodiments of this application, before performing target region recognition processing on the cathode image and the anode image using the trained target recognition algorithm to obtain the target region recognition result of the battery, the method further includes:

[0024] Cathode and anode images of different batteries are acquired, and a training dataset is constructed based on the cathode and anode images; wherein, the training dataset includes label information of the target parts in the cathode and anode images;

[0025] The initial target recognition algorithm is trained using the training dataset to obtain the trained target recognition algorithm.

[0026] In this embodiment, the initial target recognition algorithm is trained by constructing a training dataset by collecting cathode and anode images of different batteries, which can improve the robustness of the target recognition algorithm and the recognition accuracy of the trained target recognition algorithm.

[0027] In some embodiments of this application, after performing coordinate extraction processing on the recognition result to obtain the position detection result of the target part, the method further includes:

[0028] If the position detection result of the target part meets the preset offset condition, it is determined that the position of the target part has shifted; otherwise, it is determined that the position of the target part has not shifted.

[0029] In this embodiment, by setting preset offset conditions, it can be determined whether the position of the target part in the battery has shifted, thereby improving the welding accuracy during the battery soft connection process.

[0030] Secondly, embodiments of this application provide a detection device, characterized in that it includes an acquisition unit, an identification unit, and a processing unit;

[0031] The acquisition unit is used to acquire cathode and anode images of the battery;

[0032] The recognition unit is used to perform target part recognition processing on the cathode image and the anode image using a trained target recognition algorithm to obtain the target part recognition result of the battery; wherein, the trained target recognition algorithm is obtained by training an initial target recognition algorithm based on images of different batteries;

[0033] The processing unit is used to perform coordinate extraction processing on the recognition result to obtain the position detection result of the target part.

[0034] In this embodiment, the detection device can train a target recognition algorithm using images of different batteries, and then use the trained target recognition algorithm to perform target recognition on the cathode and anode images of the battery. This can be widely applied to target recognition of different types of batteries, effectively improving the anti-interference and robustness of target part detection on the battery. After obtaining the recognition result of the target part, the coordinates of the recognition result can be directly extracted to determine the location detection result of the target part to be detected. This enables accurate positioning of the target part in the battery, greatly improving the accuracy of detection.

[0035] Thirdly, embodiments of this application provide a detection device, which includes a processor and a memory storing executable instructions of the processor; when the executable instructions are executed by the processor, the above-described detection method is implemented.

[0036] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described detection method.

[0037] Fifthly, embodiments of this application provide a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps in the above-described detection method. Attached Figure Description

[0038] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiments below. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application.

[0039] Figure 1 This is a schematic diagram illustrating the implementation process of the detection method proposed in the embodiments of this application;

[0040] Figure 2 A schematic diagram of the identification results proposed in the embodiments of this application. Figure 1 ;

[0041] Figure 3 A schematic diagram of the identification results proposed in the embodiments of this application. Figure 2 ;

[0042] Figure 4 This is a schematic diagram of the composition and structure of the detection device proposed in the embodiments of this application. Figure 1 ;

[0043] Figure 5 This is a schematic diagram of the composition and structure of the detection device proposed in the embodiments of this application. Figure 2 . Detailed Implementation

[0044] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the relevant application and not for limiting the application. Furthermore, it should be noted that, for ease of description, only the parts related to the relevant application are shown in the accompanying drawings.

[0045] The application of new energy batteries in daily life and industry is becoming increasingly widespread. For example, new energy vehicles equipped with batteries are already widely used, and batteries are also increasingly being applied in energy storage. New energy batteries are not only used in energy storage systems for hydropower, thermal power, wind power, and solar power plants, but also widely used in electric vehicles such as electric bicycles, electric motorcycles, and electric cars, as well as in aerospace and other fields. With the continuous expansion of the application areas of power batteries, the market demand for them is also constantly increasing.

[0046] In the battery manufacturing process, to detect misalignment of the anode and cathode adapter plates and solder joints in the lithium battery soft connection process, it is necessary to inspect and precisely position the cover plate and adapter plates. However, in actual production, the variability of lighting, the diversity of materials, and the instability of processes can all cause significant interference to traditional positioning logic.

[0047] Current detection algorithms typically employ Hough transform and template matching, requiring template creation followed by matching. For example, precise positioning can be achieved using traditional calipers. The midpoint of the side edge is determined by the intersection of the top and bottom edges of the adapter piece and the side edge. Then, based on the actual distance from the side edge to the center of the cover plate groove, a circle is drawn to the center of the groove, achieving coarse positioning of the groove and thus determining the corresponding center of the cover plate. However, this method is relatively cumbersome, inconvenient for changeovers and production line deployment, and prone to significant positioning misalignment. When lighting fails to highlight the edges of the adapter piece and the arc contour of the cover plate, or when process changes cause positioning misalignment, it is often necessary to adjust the confidence level and region of interest. The algorithm is not compatible with parameters such as interest (ROI), resulting in low intelligent recognition performance. Furthermore, the template matching has low anti-interference capability, requiring template replacement depending on the cover plate model, leading to poor compatibility during template change. The adapter plate is covered with blue glue, which reflects light when folded, affecting the detection performance of the Hough transform and template matching combination method. In other words, the anti-interference capability of the above method is relatively low, directly causing positioning deviation and thus affecting the overkill rate.

[0048] To address the shortcomings of current detection methods, this application trains a target recognition algorithm. This trained algorithm first identifies target areas in the cathode and anode images of the battery. Compared to current methods using Hough transform and template matching, this approach removes unnecessary background interference, quickly identifies the target to be detected, and effectively improves the detection's anti-interference and robustness. Furthermore, during the Chela transformation process, only training sets need to be built for different types of batteries for automatic target area identification, eliminating the need for manual ROI setting for coarse localization. This significantly reduces deployment difficulty and improves convenience. After obtaining the target area identification results, coordinates can be directly extracted to determine the location of the target area, thus accurately locating the target within the battery and greatly improving detection accuracy.

[0049] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0050] One embodiment of this application provides a detection method, such as... Figure 1 As shown, the detection method of the detection device may include the following steps:

[0051] Step 101: Obtain the cathode and anode images of the battery.

[0052] In the embodiments of this application, the detection device may first acquire cathode and anode images of the battery.

[0053] It should be noted that, in the embodiments of this application, the battery can be assembled from one or more battery cells; the battery can be a single battery cell. A single battery cell refers to a basic unit capable of converting chemical energy into electrical energy, and can be used to manufacture battery modules or battery packs for supplying power to electrical devices. A single battery cell can be a rechargeable battery, which refers to a battery cell that can be recharged after discharge to activate its active materials and continue to be used. A single battery cell can be a lithium-ion battery, sodium-ion battery, sodium-lithium-ion battery, lithium metal battery, sodium metal battery, lithium-sulfur battery, magnesium-ion battery, nickel-metal hydride battery, nickel-cadmium battery, lead-acid battery, etc., and the embodiments of this application are not limited to this.

[0054] In embodiments of this application, the battery may also be a single physical module comprising one or more battery cells to provide higher voltage and capacity. When there are multiple battery cells, the multiple battery cells are connected in series, parallel, or mixed via a busbar.

[0055] It is understood that, in the embodiments of this application, a cathode image refers to an image of the cathode portion of the battery, and an anode image refers to an image of the anode portion of the battery; the cathode image can show the structure included in the cathode portion of the battery, and the anode image can show the structure included in the anode portion of the battery.

[0056] In the embodiments of this application, the cathode and anode portions of the battery can be photographed using any image acquisition device to obtain cathode and anode images; for example, the image acquisition device can be a CCD camera, and this application does not limit it.

[0057] Step 102: Use the trained target recognition algorithm to identify the target parts of the cathode and anode images to obtain the identification results of the target parts of the battery.

[0058] In the embodiments of this application, after acquiring the cathode image and anode image of the battery, the detection device can use the trained target recognition algorithm to perform target part recognition processing on the cathode image and anode image to obtain the target part recognition result of the battery.

[0059] It should be noted that, in the embodiments of this application, the target part refers to the part of the battery that needs to be detected and identified. The target part can be at least one part of the battery, and this application does not limit it.

[0060] In some embodiments of this application, the target area may include at least one of an adapter plate and a cover plate.

[0061] It should be noted that, in the embodiments of this application, the adapter refers to a component used to connect the internal electrodes of the battery and the external circuit. It is usually made of conductive materials, such as copper foil or aluminum foil, and can provide a current conduction path so that the positive and negative electrodes of the battery can be connected to the external circuit. The adapter is located between the positive and negative electrodes in the battery structure and mainly plays the role of conducting electricity and transmitting electrons.

[0062] It should be noted that, in the embodiments of this application, the cover plate is a component used to isolate the positive and negative electrodes, and is usually made of a microporous film or a porous film, which allows lithium ions to pass through while preventing direct contact between the electrodes.

[0063] It should be noted that, in the embodiments of this application, the target recognition algorithm can be any algorithm that can be used for target recognition, and the specific algorithm is not limited in this application.

[0064] In some embodiments of this application, the target recognition algorithm may employ the YOLOv5 algorithm, which can achieve real-time target detection or recognition through a single neural network model, while being optimized in terms of accuracy and speed.

[0065] It should be noted that, in the embodiments of this application, the trained target recognition algorithm is obtained by training the initial target recognition algorithm based on images of different batteries, and the model structure and parameters of the initial target recognition algorithm are not limited in this application.

[0066] For example, the initial target recognition algorithm is the YOLOv5-seg algorithm, whose structure can be divided into a main model and some additional modules. The main model can use CSPDarknet53 as the backbone network, which is a lightweight network structure with high performance and efficiency. Some enhancement modules and techniques can be used as additional modules, such as using Spatial Pyramid Pooling (SPP) to increase the size of the receptive field and improve the ability to detect targets.

[0067] It is understood that, in the embodiments of this application, the recognition result can display the identified target area. For example, in the recognition result, a mask can be used to indicate the adapter plate and cover plate in the cathode image; wherein, the mask is a binary image in which the pixel value is 0 or 1, the area with a pixel value of 1 indicates the area that should be processed, and the area with a pixel value of 0 indicates the area that does not need to be processed; for example, in the recognition result, the pixel value of the area of ​​the adapter plate in the cathode image is 1, and the recognition result also includes label information identifying this area as the adapter plate.

[0068] In the embodiments of this application, when the trained target recognition algorithm performs target part recognition processing on cathode and anode images, it can determine the recognition result based on the confidence level. For example, if the confidence level of a certain area in the anode image being matched as a cover plate is 100, then the area is indicated by a mask and the label information of the cover plate is generated, thereby generating the recognition result.

[0069] It should be noted that, in the embodiments of this application, during the process of identifying target parts in cathode and anode images using the trained target recognition algorithm, after the target parts are identified, the target parts in the cathode and anode images can be segmented. For example, after identifying the cover plate and the adapter plate in the cathode image, the cover plate area and the adapter plate area in the image can be segmented pixel by pixel to generate recognition results that indicate the cover plate and adapter plate areas respectively.

[0070] For example, the recognition result obtained by processing the anode image of the battery can be as follows: Figure 2 As shown, the circular area is the circular area 111 corresponding to the pad of the adapter piece identified in the anode image; the area below the adapter piece is the area 112 corresponding to the cover plate identified in the anode image; the recognition result obtained by processing the cathode image of the battery can be as follows: Figure 3 As shown, the circular area is the circular area 113 corresponding to the pad of the adapter piece identified in the cathode image; the area below the adapter piece is the area 114 corresponding to the cover plate identified in the cathode image.

[0071] Step 103: Perform coordinate extraction processing on the recognition results to obtain the position detection results of the target part.

[0072] In the embodiments of this application, after the detection device uses the trained target recognition algorithm to perform target part recognition processing on the cathode image and anode image to obtain the target part recognition result of the battery, it can perform coordinate extraction processing on the recognition result to obtain the position detection result of the target part.

[0073] In some embodiments of this application, when the target part is at least one of a transition piece and a cover plate, when the detection device performs coordinate extraction processing on the recognition result to obtain the position detection result of the target part, it can perform coordinate extraction processing on the center point of the target area based on the recognition result of the transition piece to obtain the position detection result of the transition piece; and perform coordinate extraction processing on the edge line of the cover plate based on the recognition result of the cover plate to obtain the position detection result of the cover plate.

[0074] It should be noted that, in the embodiments of this application, the target area is the circular area corresponding to the pad in the adapter piece, and the position detection result of the adapter piece includes the position detection result of the center point of the pad.

[0075] In some embodiments of this application, the identification result of the adapter piece may include the target area indicated by a mask, that is, the circular area corresponding to the pad indicated by the mask.

[0076] In some embodiments of this application, when the detection device extracts the coordinates of the center point of the target area based on the identification result of the adapter piece to obtain the position detection result of the adapter piece, it can perform a circular area fitting process based on the first contour point in the identification result of the adapter piece to obtain a fitted circular area; and determine the center coordinates of the fitted circular area, and use the center coordinates as the position detection result of the center point of the pad.

[0077] It should be noted that, in the embodiments of this application, the first contour point represents the contour point of the circular area corresponding to the pad in the adapter piece obtained from the identification result of the adapter piece.

[0078] It is understood that, in the embodiments of this application, the fitted circular region represents the result of fitting the region of the circular pad.

[0079] In some embodiments of this application, when the detection device performs circular region fitting processing based on the first contour point in the recognition result of the adapter piece to obtain a fitted circular region, it can perform circular region fitting processing based on the coordinates of the first contour point in the recognition result of the adapter piece to obtain an initial fitted arc; then calculate the distance information between the first contour point and the fitted arc; and then, if the distance information does not meet the preset conditions, perform weighted calculation on the coordinates of the first contour point to obtain the filtered first contour point; thereby determining the fitted circular region based on the filtered first contour point.

[0080] In the embodiments of this application, if the distance information meets the preset conditions, the fitted circular region can be determined based on the initial fitted arc.

[0081] It should be noted that, in the embodiments of this application, the preset condition can be a preset number of distance information that is less than or equal to a preset distance threshold; wherein, the specific values ​​of the preset number and the preset distance threshold are not limited in this application.

[0082] It should be noted that, in the embodiments of this application, the initial fitted arc represents the edge of the circular region obtained by the initial fitting.

[0083] For example, the preset condition is that the distance information between 90% of the first contour points and the fitted arc is less than or equal to x. Assuming there are 100 first contour points, if the distance information between 91 of the first contour points and the fitted arc is less than or equal to x, then the distance information can be determined to meet the preset condition. However, if only 80 of the first contour points and the fitted arc have a distance information less than or equal to x, then the distance information can be determined to not meet the preset condition, and a weighted calculation needs to be performed on the 100 first contour points.

[0084] It should be noted that, in the embodiments of this application, the purpose of weighting the coordinates of the first contour points is to reduce the influence of these first contour points with excessively large distance information, so as to improve the accuracy of the fitting.

[0085] It should be noted that, in the embodiments of this application, when performing weighted calculations on the coordinates of the first contour points, the weight assigned to the coordinates of each first contour point can be set by the applicant and is not limited in this application.

[0086] In some embodiments of this application, when performing weighted calculations on the coordinates of the first contour points to obtain the filtered first contour points, outliers in the first contour points can be removed based on the weighted average obtained from the weighted calculation of the coordinates of the first contour points, thereby determining the remaining points as the filtered first contour points; for example, points whose absolute value of the difference between the coordinates of the first contour points and the weighted average is greater than a first preset value can be regarded as outliers, and the value of the first preset value is not limited in this application.

[0087] For example, given n first contour points, when performing a weighted calculation on these n first contour points, the weighted average can be expressed as (weight 1 × point 1 + weight 2 × point 2 + ... + weight n × point n) ÷ (weight 1 + weight 2 + ... + weight n), where point 1 to point n represent the coordinates of the n first contour points, and weight 1 to weight n represent the weights assigned to the n first contour points. The weighted average can then be used to determine the filtered first contour points from among the n first contour points.

[0088] In some embodiments of this application, when the detection device extracts the coordinates of the cover edge line based on the cover recognition result to obtain the cover position detection result, it can sort the ordinate values ​​of the second contour points in the cover recognition result in descending order to obtain the sorting result; then, based on the sorting result, it determines a first number of third contour points and a first number of fourth contour points; thereby obtaining the cover position detection result based on the third contour points and the fourth contour points.

[0089] It should be noted that, in the embodiments of this application, the second contour point represents the contour point of the cover plate area determined from the identification result of the cover plate.

[0090] It is understood that, in the embodiments of this application, the sorting result includes the ordinate values ​​of the second contour points arranged from largest to smallest.

[0091] It should be noted that, in the embodiments of this application, the third contour point is determined based on the second contour points corresponding to the first number of ordinate values ​​in the sorting result, starting from the largest ordinate value; the fourth contour point is determined based on the second contour points corresponding to the first number of ordinate values ​​in the sorting result, starting from the smallest ordinate value.

[0092] It should also be noted that, in the embodiments of this application, the value of the first quantity is not limited. For example, the first quantity can be 10%.

[0093] For example, the second contour point includes 10 points, whose respective ordinate values ​​are 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10; the sorting result after being arranged in descending order is 10, 9, 8, 7, 6, 5, 4, 3, 2, and 1; assuming the first quantity is 20%, the third contour point can be determined to include the second contour point corresponding to the ordinate value 10 and the second contour point corresponding to the ordinate value 9; the fourth contour point can include the second contour point corresponding to the ordinate value 1 and the second contour point corresponding to the ordinate value 2.

[0094] It is understood that, in the embodiments of this application, the minimum ordinate value of the third contour point is greater than the maximum ordinate value of the fourth contour point.

[0095] In some embodiments of this application, the position detection results of the cover plate include the position detection results of the upper edge and the lower edge of the cover plate.

[0096] In some embodiments of this application, when the detection device obtains the position detection result of the cover plate based on the third contour point and the fourth contour point, it can calculate the average value of each ordinate value in the third contour point and the average value of each ordinate value in the fourth contour point to obtain the first average value corresponding to the third contour point and the second average value corresponding to the fourth contour point; determine the abnormal coordinate points in the third contour point based on each ordinate value in the third contour point and the first average value, and determine the abnormal coordinate points in the fourth contour point based on each ordinate value in the fourth contour point and the second average value; perform a rejection process on the abnormal coordinate points in the third contour point and the abnormal coordinate points in the fourth contour point to obtain the filtered third contour point and the filtered fourth contour point; determine the upper edge position detection result based on the filtered third contour point, and determine the lower edge position detection result based on the filtered fourth contour point.

[0097] For example, the ordinate values ​​of the third contour point are represented as a, b, c, and the ordinate values ​​of the fourth contour point are represented as d, e, f; the first average value can be (a+b+c)÷3, and the second average value can be (d+e+f)÷3.

[0098] In some embodiments of this application, when determining abnormal coordinate points in the third contour points based on each ordinate value and the first average value, the third contour points whose absolute value of the difference between their ordinate values ​​and the first average value is greater than a second preset value can be determined as abnormal coordinate points in the third contour points. Correspondingly, when determining abnormal coordinate points in the fourth contour points based on each ordinate value and the second average value, the fourth contour points whose absolute value of the difference between their ordinate values ​​and the second average value is greater than a third preset value can be determined as abnormal coordinate points in the fourth contour points. The second and third preset values ​​are not limited in this application.

[0099] For example, the second preset value is denoted as A. The difference between each ordinate value a, b, c in the third contour point and its corresponding first average value is calculated, and the absolute value is taken to obtain a1, b1, c1, where b1 > A. Then, the third contour point corresponding to the ordinate value b is determined as an abnormal coordinate point and removed. The third preset value is denoted as B. The difference between each ordinate value d, e, f in the fourth contour point and its corresponding second average value is calculated, and the absolute value is taken to obtain d1, e1, f1, where f1 > B. Then, the fourth contour point corresponding to the ordinate value f is determined as an abnormal coordinate point and removed.

[0100] In some embodiments of this application, since the cover plate is rectangular in shape and includes four sides, if the cover plate is rotated by 90°, the matching upper and lower edges will change. At this time, the upper edge position detection result and the lower edge position detection result in this direction can also be determined in the manner described above for determining the position detection result of the cover plate.

[0101] In the embodiments of this application, before the detection device performs target region recognition processing on the cathode image and anode image using the trained target recognition algorithm to obtain the target region recognition result of the battery, i.e. before step 102, the following steps may also be included:

[0102] Step 104: Collect cathode and anode images of different batteries, and construct a training dataset based on the cathode and anode images.

[0103] In the embodiments of this application, before the detection device performs target part recognition processing on the cathode image and anode image using the trained target recognition algorithm to obtain the target part recognition result of the battery, it can collect cathode images and anode images of different batteries and construct a training dataset based on the cathode image and anode image.

[0104] It should be noted that, in the embodiments of this application, the training dataset may include label information of target parts in cathode and anode images of different batteries.

[0105] In some embodiments of this application, the training dataset may include label information for the cover plate and adapter plate regions in cathode and anode images of different batteries.

[0106] Step 105: Train the initial target recognition algorithm using the training dataset to obtain the trained target recognition algorithm.

[0107] In the embodiments of this application, after the detection device acquires cathode and anode images of different batteries and constructs a training dataset based on the cathode and anode images, it can use the training dataset to train the initial target recognition algorithm to obtain the trained target recognition algorithm.

[0108] In some embodiments of this application, the initial target recognition algorithm may be the YOLOv5 algorithm.

[0109] In some embodiments of this application, when the detection device trains an initial target recognition algorithm using a training dataset to obtain a trained target recognition algorithm, it can use the initial target recognition algorithm to identify target parts in the training dataset to obtain a first result; calculate a first loss function value between the first result and the label information; update the parameters in the initial target recognition algorithm based on the first loss function value to obtain an updated target recognition algorithm; and thus determine the trained target recognition algorithm based on the updated target recognition algorithm.

[0110] In some embodiments of this application, when the detection device determines the trained target recognition algorithm based on the updated target recognition algorithm, it can use the updated target recognition algorithm to perform target part recognition processing on the training dataset to obtain a second result; calculate the second loss function value between the second result and the label information; if the second loss function value is less than or equal to a preset value, the updated target recognition algorithm is determined as the trained target recognition algorithm; otherwise, the updated target recognition algorithm is updated based on the second loss function value to obtain a first target recognition algorithm, until the third loss function value obtained based on the first target recognition algorithm is less than or equal to the preset value, at which point the first target recognition algorithm is determined as the trained target recognition algorithm.

[0111] In some embodiments of this application, the first result may include the result obtained by identifying and detecting the cover plate and adapter piece in the training dataset using an initial target recognition algorithm.

[0112] It is understood that, in the embodiments of this application, the loss function value can measure the difference between the output of the target recognition algorithm and the real label information. The smaller the loss function value, the closer the output of the target recognition algorithm is to the real label information. The loss function is used to continuously adjust the parameters in the target recognition algorithm to minimize the error, so that the target recognition algorithm can better fit the training data and improve the generalization ability. This application does not limit the loss function. For example, the loss function can be the mean squared error (MSE), cross-entropy error, and log loss.

[0113] It is also understood that, in the embodiments of this application, when the loss function value is less than or equal to a preset value, it indicates that the difference between the result of the target recognition algorithm input in the current iteration and the real label information has reached the requirement and the convergence condition is met, and training can be stopped to obtain the trained target recognition algorithm; wherein, the value of the preset value is not limited in this application.

[0114] In the embodiments of this application, after the detection device performs coordinate extraction processing on the recognition result to obtain the position detection result of the target part, i.e. after step 103, it may further include the following steps:

[0115] Step 106: If the position detection result of the target part meets the preset offset conditions, determine that the position of the target part has shifted; otherwise, determine that the position of the target part has not shifted.

[0116] In the embodiments of this application, after the detection device performs coordinate extraction processing on the recognition result to obtain the position detection result of the target part, it can determine that the position of the target part has shifted if the position detection result of the target part meets the preset offset condition; otherwise, it determines that the position of the target part has not shifted.

[0117] It should be noted that, in the embodiments of this application, the preset offset condition represents the judgment condition used to determine whether the position of the target part has shifted.

[0118] In some embodiments of this application, the preset offset condition can be that the center point of the pad in the adapter piece is located on the center line of the cover plate; or it can be that the offset angle of the cover plate is less than or equal to the preset offset amount.

[0119] In some embodiments of this application, when the center point of the adapter pad is determined to be on the center line of the cover plate based on the detection results of the upper edge position and the lower edge position of the cover plate and the detection results of the position of the adapter piece, it can be determined that the cover plate and the adapter piece are not offset; otherwise, it is determined that the cover plate and the adapter piece are offset.

[0120] In some embodiments of this application, the centerline of the cover plate can be determined based on the detection results of the upper edge position and the lower edge position of the cover plate, thereby determining whether the center point of the adapter pad is located on the centerline of the cover plate.

[0121] In some embodiments of this application, when the offset angle of the cover plate is determined to be less than or equal to a preset offset amount based on the detection results of the upper edge position and the lower edge position of the cover plate and the preset direction, it can be determined that the cover plate and the adapter piece are not offset; otherwise, it is determined that the cover plate and the adapter piece are offset.

[0122] It should be noted that, in the embodiments of this application, the preset direction can be used to measure whether the cover plate has shifted; the preset direction can be determined based on the position detection result of the adapter piece, so that the degree of shift of the cover plate relative to the adapter piece can be measured based on the preset direction; the preset direction can also be a pre-set direction.

[0123] It is understood that in the embodiments of this application, the preset offset is a value used to measure the offset angle of the cover plate, and its specific value is not limited in this application; for example, when the offset angle of the cover plate is determined to be 15° and the preset offset is 5° based on the detection results of the upper edge position and the lower edge position of the cover plate and the preset direction, it can be determined that the cover plate has shifted because the offset angle of the cover plate is greater than the preset offset.

[0124] For example, based on the position detection results of the target part, it can be determined whether the target part has shifted; such as Figure 2 As shown, the position detection result of the adapter piece is the position detection result of the center point of the pad. The position detection result of the center point of the pad is displayed as a circular area. The position detection results of the upper and lower edges of the cover plate, as well as the offset angle of the cover plate, are represented by vertical lines with arrows after processing. The vertical line located between the two vertical lines can represent the center line of the cover plate. For example... Figure 2 As shown, it can be assumed that the center point of the pad is not located on the center line of the cover plate, and there is an offset. At the same time, it can also be determined that the offset angle is greater than the preset offset amount based on the position detection results of any vertical line with an arrow, that is, any edge of the cover plate, thus confirming the existence of an offset.

[0125] In the embodiments of this application, the detection device can acquire cathode and anode images of the battery; use a trained target recognition algorithm to perform target part recognition processing on the cathode and anode images to obtain the target part recognition result of the battery; and perform coordinate extraction processing on the recognition result to obtain the position detection result of the target part. Therefore, the detection device in this application can effectively improve the anti-interference and robustness of the detection by training the target recognition algorithm to first perform target recognition on the cathode and anode images of the battery; after obtaining the target part recognition result, coordinate extraction can be directly performed on the recognition result to determine the position detection result of the target part to be detected, thereby accurately locating the target part in the battery and greatly improving the detection accuracy.

[0126] Based on the above embodiments, in another embodiment of this application, a detection device is provided, such as... Figure 4 As shown, the detection device 1 may include an acquisition unit 11, an identification unit 12, a processing unit 13, a construction unit 14, a training unit 15, and a determination unit 16.

[0127] The acquisition unit is used to acquire cathode and anode images of the battery.

[0128] The recognition unit is used to perform target part recognition processing on the cathode image and anode image using the trained target recognition algorithm to obtain the target part recognition result of the battery; wherein, the trained target recognition algorithm is obtained by training the initial target recognition algorithm based on images of different batteries.

[0129] The processing unit is used to extract coordinates from the recognition results to obtain the location detection results of the target part.

[0130] In some embodiments, the processing unit can also be used to extract the coordinates of the center point of the target area based on the recognition result of the adapter piece to obtain the position detection result of the adapter piece; and to extract the coordinates of the edge line of the cover plate based on the recognition result of the cover plate to obtain the position detection result of the cover plate.

[0131] In some embodiments, the processing unit can also be used to perform a circular region fitting process based on the first contour point in the identification result of the adapter piece to obtain a fitted circular region; and to determine the center coordinates in the fitted circular region and use the center coordinates as the position detection result of the center point of the pad.

[0132] In some embodiments, the processing unit can also be used to perform circular region fitting processing based on the coordinates of the first contour point in the recognition result of the adapter piece to obtain an initial fitted arc; and to calculate the distance information between the first contour point and the fitted arc; and to perform weighted calculation on the coordinates of the first contour point if the distance information does not meet the preset conditions, so as to obtain the filtered first contour point; and to determine the fitted circular region based on the filtered first contour point.

[0133] In some embodiments, the processing unit may further be configured to sort the ordinate values ​​of the second contour points in the identification result of the cover plate in descending order to obtain a sorting result; and determine a first number of third contour points and a first number of fourth contour points based on the sorting result; wherein the minimum ordinate value of the third contour points is greater than the maximum ordinate value of the fourth contour points; and obtain the position detection result of the cover plate based on the third contour points and the fourth contour points.

[0134] In some embodiments, the processing unit may further be configured to calculate the average value of each ordinate value in the third contour point and each ordinate value in the fourth contour point to obtain a first average value corresponding to the third contour point and a second average value corresponding to the fourth contour point; and to determine abnormal coordinate points in the third contour point based on each ordinate value in the third contour point and the first average value, and to determine abnormal coordinate points in the fourth contour point based on each ordinate value in the fourth contour point and the second average value; and to perform a rejection process on the abnormal coordinate points in the third contour point and the abnormal coordinate points in the fourth contour point to obtain filtered third contour points and filtered fourth contour points; and to determine the upper edge position detection result based on the filtered third contour points and the lower edge position detection result based on the filtered fourth contour points.

[0135] The construction unit 14 can be used to collect cathode and anode images of different batteries and construct a training dataset based on the cathode and anode images before the recognition unit uses the trained target recognition algorithm to perform target part recognition processing on the cathode and anode images to obtain the target part recognition result of the battery; wherein, the training dataset includes the label information of the target parts in the cathode and anode images.

[0136] Training unit 15 can be used to train the initial target recognition algorithm using the training dataset to obtain the trained target recognition algorithm.

[0137] The training unit 15 can also be used to identify target parts in the training dataset using the initial target recognition algorithm to obtain a first result; calculate a first loss function value between the first result and the label information; update the parameters in the initial target recognition algorithm based on the first loss function value to obtain an updated target recognition algorithm; and determine the trained target recognition algorithm based on the updated target recognition algorithm.

[0138] The training unit 15 can also be used to perform target part recognition processing on the training dataset using the updated target recognition algorithm to obtain a second result; and to calculate a second loss function value between the second result and the label information. If the second loss function value is less than or equal to a preset value, the updated target recognition algorithm is determined as the trained target recognition algorithm. Otherwise, the updated target recognition algorithm is updated based on the second loss function value to obtain a first target recognition algorithm, until the third loss function value obtained based on the first target recognition algorithm is less than or equal to a preset value, thus obtaining the trained target recognition algorithm.

[0139] The determining unit 16 can be used to determine that the position of the target part has shifted if the position detection result of the target part meets the preset offset condition after the processing unit performs coordinate extraction processing on the recognition result and obtains the position detection result of the target part; otherwise, it can determine that the position of the target part has not shifted.

[0140] Based on the above embodiments, in another embodiment of this application... Figure 5 This is a schematic diagram of the composition and structure of the detection device proposed in the embodiments of this application. Figure 2 ,like Figure 5 As shown, the detection device 1 proposed in this application embodiment may further include a processor 17 and a memory 18 storing instructions executable by the processor 17; further, the detection device 1 may further include a communication interface 19 and a bus 110 for connecting the processor 17, the memory 18 and the communication interface 19.

[0141] In the embodiments of this application, the processor 17 can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, and microprocessor. It is understood that for different devices, the electronic device used to implement the above-mentioned processor function can also be other types, and this application embodiment does not specifically limit this. The detection device 1 may also include a memory 18, which can be connected to the processor 17. The memory 18 is used to store executable program code, which includes computer operation instructions. The memory 18 may include high-speed RAM memory and may also include non-volatile memory, such as at least two disk drives.

[0142] In embodiments of this application, bus 110 is used to connect communication interface 19, processor 17, and memory 18, as well as the mutual communication between these devices.

[0143] In embodiments of this application, memory 18 is used to store instructions and data.

[0144] Furthermore, in the embodiments of this application, the processor 17 is used to acquire cathode and anode images of the battery;

[0145] The trained target recognition algorithm is used to identify target parts in cathode and anode images to obtain the target part identification results of the battery; the trained target recognition algorithm is obtained by training the initial target recognition algorithm based on images of different batteries.

[0146] The recognition results are processed by coordinate extraction to obtain the position detection results of the target part.

[0147] In practical applications, the aforementioned memory 18 can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or a combination of the above types of memory, and provide instructions and data to the processor 17.

[0148] Furthermore, in this embodiment, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0149] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0150] This application provides a detection device that can acquire cathode and anode images of a battery; use a trained target recognition algorithm to identify target parts in the cathode and anode images to obtain the identification result of the target part of the battery; and perform coordinate extraction processing on the identification result to obtain the position detection result of the target part. Therefore, the detection device can effectively improve the anti-interference and robustness of the detection by training the target recognition algorithm to first identify targets in the cathode and anode images of the battery. After obtaining the identification result of the target part, coordinate extraction can be directly performed on the identification result to determine the position detection result of the target part to be detected, thereby accurately locating the target part in the battery and greatly improving the accuracy of the detection.

[0151] Specifically, the program instructions corresponding to a detection method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to a detection method in the storage media are read or executed by an electronic device, the following steps are included:

[0152] Acquire images of the battery's cathode and anode;

[0153] The trained target recognition algorithm is used to identify target parts in cathode and anode images to obtain the target part identification results of the battery; the trained target recognition algorithm is obtained by training the initial target recognition algorithm based on images of different batteries.

[0154] The recognition results are processed by coordinate extraction to obtain the position detection results of the target part.

[0155] This application provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps in the method provided in the above-described method embodiments.

[0156] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects.

[0157] This application is described with reference to schematic and / or block diagrams of implementations of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the schematic and / or block diagrams can be implemented by computer program instructions, and combinations of blocks in the schematic and / or block diagrams can be implemented. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the schematic and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0158] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in the implementation flow diagram. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0159] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0160] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application.

Claims

1. A detection method, characterized in that, The method includes: Acquire images of the battery's cathode and anode; The trained target recognition algorithm is used to identify target parts in the cathode image and the anode image to obtain the target part identification result of the battery; wherein, the trained target recognition algorithm is obtained by training an initial target recognition algorithm based on images of different batteries; The recognition results are processed by coordinate extraction to obtain the position detection results of the target part.

2. The detection method according to claim 1, characterized in that, The target location includes at least one of a transition piece and a cover plate; the coordinate extraction processing of the identification result to obtain the position detection result of the target location includes: Based on the recognition result of the adapter piece, the coordinates of the center point of the target area are extracted to obtain the position detection result of the adapter piece; Based on the recognition result of the cover plate, the coordinates of the cover plate edge line are extracted to obtain the position detection result of the cover plate.

3. The detection method according to claim 2, characterized in that, The target area is the circular area corresponding to the pads in the adapter piece, and the position detection result of the adapter piece includes the position detection result of the center point of the pads; the step of extracting the coordinates of the center point of the target area based on the identification result of the adapter piece to obtain the position detection result of the adapter piece includes: The circular region is fitted based on the first contour point in the recognition result of the adapter piece to obtain the fitted circular region. Determine the center coordinates of the fitted circular region, and use the center coordinates as the position detection result of the center point of the pad.

4. The detection method according to claim 3, characterized in that, The step of fitting the circular region based on the first contour points in the recognition result of the adapter piece to obtain the fitted circular region includes: Based on the first contour point in the recognition result of the adapter piece, a circular region is fitted to obtain an initial fitted circular arc. Calculate the distance information between the first contour point and the fitted circular arc; If the distance information does not meet the preset conditions, the coordinates of the first contour point are weighted and calculated to obtain the filtered first contour point. The fitted circular region is determined based on the first contour points after filtering.

5. The detection method according to claim 2, characterized in that, The step of extracting the coordinates of the cover plate edge lines based on the recognition results of the cover plate to obtain the position detection results of the cover plate includes: The ordinate values ​​of the second contour points in the identification results of the cover plate are sorted in descending order to obtain the sorting result; Based on the sorting result, a first number of third contour points and a first number of fourth contour points are determined; wherein, the minimum ordinate value of the third contour point is greater than the maximum ordinate value of the fourth contour point; The position detection result of the cover plate is obtained based on the third contour point and the fourth contour point.

6. The detection method according to claim 5, characterized in that, The position detection result of the cover plate includes the position detection result of the upper edge and the position detection result of the lower edge of the cover plate; obtaining the position detection result of the cover plate based on the third contour point and the fourth contour point includes: The average values ​​of each ordinate value in the third contour point and each ordinate value in the fourth contour point are calculated to obtain the first average value corresponding to the third contour point and the second average value corresponding to the fourth contour point. Abnormal coordinate points in the third contour points are determined based on each ordinate value in the third contour points and the first average value; abnormal coordinate points in the fourth contour points are determined based on each ordinate value in the fourth contour points and the second average value. Abnormal coordinate points in the third contour points and abnormal coordinate points in the fourth contour points are removed respectively to obtain the filtered third contour points and the filtered fourth contour points. The upper edge position detection result is determined based on the filtered third contour point, and the lower edge position detection result is determined based on the filtered fourth contour point.

7. The detection method according to any one of claims 1 to 6, characterized in that, Before using the trained target recognition algorithm to identify target parts in the cathode image and the anode image to obtain the target part identification result of the battery, the method further includes: Cathode and anode images of different batteries are collected, and a training dataset is constructed based on the cathode and anode images; wherein, the training dataset includes label information of the target parts in the cathode and anode images; The initial target recognition algorithm is trained using the training dataset to obtain the trained target recognition algorithm.

8. The detection method according to claim 7, characterized in that, The step of training the initial target recognition algorithm using the training dataset to obtain the trained target recognition algorithm includes: The target parts in the training dataset are identified using the initial target recognition algorithm to obtain a first result; Calculate the first loss function value between the first result and the label information; The parameters in the initial target recognition algorithm are updated based on the first loss function value to obtain the updated target recognition algorithm; The trained target recognition algorithm is determined based on the updated target recognition algorithm.

9. The detection method according to claim 8, characterized in that, Determining the trained target recognition algorithm based on the updated target recognition algorithm includes: The updated target recognition algorithm is used to identify the target parts in the training dataset to obtain a second result; Calculate the second loss function value between the second result and the label information. If the second loss function value is less than or equal to a preset value, determine the updated target recognition algorithm as the trained target recognition algorithm. Otherwise, update the updated target recognition algorithm based on the second loss function value to obtain the first target recognition algorithm. Continue until the third loss function value obtained based on the first target recognition algorithm is less than or equal to the preset value, then determine the first target recognition algorithm as the trained target recognition algorithm.

10. The detection method according to claim 7, characterized in that, After performing coordinate extraction processing on the recognition result to obtain the position detection result of the target part, the method further includes: If the position detection result of the target part meets the preset offset condition, it is determined that the position of the target part has shifted; otherwise, it is determined that the position of the target part has not shifted.

11. A detection device, characterized in that, It includes an acquisition unit, an identification unit, and a processing unit; The acquisition unit is used to acquire cathode and anode images of the battery; The recognition unit is used to perform target part recognition processing on the cathode image and the anode image using a trained target recognition algorithm to obtain the target part recognition result of the battery; wherein, the trained target recognition algorithm is obtained by training an initial target recognition algorithm based on images of different batteries; The processing unit is used to perform coordinate extraction processing on the recognition result to obtain the position detection result of the target part.

12. A detection device, characterized in that, The detection device includes a processor and a memory storing processor-executable instructions; when the executable instructions are executed by the processor, the method described in any one of claims 1 to 10 is implemented.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 10.

14. A computer program product comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the detection method according to any one of claims 1 to 10.

Images