An airtight detection method and system for an electrical connection panel of an energy storage battery upper cover

Abstract

This invention discloses a method and system for detecting the airtightness of electrical connection panels on the top cover of energy storage batteries. It aims to solve the core problems in existing testing of electrical connection panels for large-size energy storage batteries, such as false leaks caused by pressure deformation, detection distortion due to poor sealing of electrical interfaces, and low accuracy in detecting micro-leakage. The method includes: sealing and fixing the edge of the battery top cover to the electrical connection port to form a sealed testing cavity; providing adaptive surface support for the large flat surface of the panel to suppress pressure deformation; filling the cavity with a testing medium and maintaining pressure, collecting the panel deformation in real time and dynamically compensating for the total volume of the cavity in reverse to ensure a constant total volume throughout the pressure holding phase; and finally determining airtightness based on the pressure change in the cavity. This method completely eliminates false leaks, significantly improves the accuracy of micro-leakage detection, and is suitable for mass production testing requirements.

Description

Technical Field

[0001] This invention relates to the field of battery packaging airtightness testing, and more particularly to a method and system for airtightness testing of the electrical connection panel of an energy storage battery cover. Background Technology

[0002] With the rapid development of the new energy industry, the safety and reliability of energy storage batteries have become a core concern in the industry. As a key component connecting the battery pack to external circuits, the electrical connection panel on the battery cover must simultaneously achieve the dual functions of electrical conductivity and sealing protection. Its airtightness directly affects the prevention of intrusion of external contaminants such as moisture and dust, thereby avoiding serious safety accidents such as insulation failure, internal short circuits, and even thermal runaway. Currently, a significant proportion of safety accidents involving new energy storage equipment are related to battery packaging leakage. Some high-end application scenarios have even higher requirements for ultra-micro leakage detection, posing a severe challenge to the accuracy and efficiency of detection technologies.

[0003] In existing technologies, commonly used airtightness testing methods in the industry include differential pressure testing, helium mass spectrometry, traditional thermal imaging testing, and water testing. While these methods can qualitatively determine basic airtightness performance, they have significant limitations when applied to the electrical connection panels of energy storage battery covers. Differential pressure testing can only determine whether the panel is leaking overall by measuring pressure changes, but it cannot provide specific location information of the leak point, resulting in a lack of targeted rework. Helium mass spectrometry is limited by the high cost and low detection efficiency of helium, making it difficult to meet the needs of large-scale full inspection on production lines, and is usually only used for sampling inspection. Traditional thermal imaging testing, when applied to panels made of metal materials such as aluminum alloy and stainless steel, is easily interfered with by the high reflectivity and thermal radiation characteristics of the material surface, generating a large number of false signals, resulting in a high false alarm rate and poor stability. Water testing may introduce moisture residue, affecting the internal environment of the battery. In addition, none of the existing methods can achieve precise location of the leak point, making it difficult to quickly identify the root cause of defects during the production process, resulting in low quality control efficiency, increased rework costs, and limiting the overall reliability and production yield of energy storage battery products. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for airtightness detection of the electrical connection panel of the top cover of an energy storage battery, which has the advantages of accurately locating the leakage point, improving detection efficiency and product quality.

[0005] To solve the above-mentioned technical problems, the present invention provides a method for airtightness testing of the electrical connection panel of an energy storage battery cover, comprising the following steps:

[0006] S1: The panel to be tested is clamped and sealed to form a closed testing cavity. Gas is injected into the closed testing cavity, and the overall airtightness is initially inspected by detecting the pressure change in the closed testing cavity to distinguish between qualified and unqualified parts. The method is characterized by further including the following steps: S2: Preload the three-dimensional electronic model corresponding to the panel under test, complete the coordinate calibration of at least two sets of orthogonally arranged schlieren imaging systems and the detection space, and establish a unified coordinate mapping relationship between the three-dimensional electronic model, the schlieren imaging system and the detection space. S3: Carbon dioxide gas is introduced into the sealed detection chamber as a tracer gas. For the unqualified parts determined by the initial air tightness inspection, infrared thermal imaging matching the characteristic absorption band of carbon dioxide is used to image the entire area of ​​the panel under test and extract the suspected leakage area. S4: For the suspected leakage area, control the orthogonally deployed schlieren imaging system to simultaneously acquire schlieren images of the leaking jet from each orthogonal viewpoint for the suspected leakage area; S5: Extract the leakage jet trajectory from the schlieren images of each orthogonal viewpoint, obtain the jet source coordinates of the leakage jet trajectory under each viewpoint through reverse source tracing fitting, and obtain the three-dimensional spatial coordinates of the leakage point through spatial intersection fitting of each orthogonal viewpoint. S6: Map the three-dimensional spatial coordinates of the leak point to the three-dimensional electronic model described in step S2 to complete the three-dimensional annotation of the leak point.

[0007] By adopting the above technical solution, and combining the rapid full-area screening capability of infrared thermal imaging with the precise positioning capability of orthogonal schlieren imaging, the problem of traditional detection methods being unable to accurately locate leak points, having low detection efficiency, and being prone to false alarms in the airtightness detection of electrical connection panels on the top cover of energy storage batteries is effectively solved. This method can realize the three-dimensional spatial coordinate positioning and stereoscopic marking of ultra-micro leak points, providing precise guidance for subsequent rework and process optimization, thereby improving the production quality and reliability of energy storage battery panels.

[0008] The present invention is further configured such that: in step S1, carbon dioxide gas is introduced into the gas filling the sealed testing chamber for the overall airtightness preliminary inspection; in step S3, for the unqualified parts determined by the airtightness preliminary inspection, the gas pressure introduced into the sealed testing chamber is higher than the gas pressure introduced in S1.

[0009] By adopting the above technical solution, carbon dioxide gas is used as the filling gas in the initial airtightness inspection stage. This application achieves the uniformity of the detection gas throughout the entire process, avoiding the complex operation and potential time waste of switching between different gases. This not only simplifies the inspection process and improves the inspection efficiency, but also ensures the consistency of the detection medium, providing a stable foundation for subsequent refined inspection. Furthermore, for parts that fail the initial inspection, when performing infrared thermal imaging inspection, the pressure of carbon dioxide gas in the sealed detection chamber is increased, which significantly increases the pressure difference inside and outside the leak point. This increase in pressure difference makes the leak jet stronger and more obvious, thereby greatly enhancing the detection sensitivity of infrared thermal imaging for small leaks. This allows it to more effectively capture weak leak signals and extract more accurate suspected leak areas. This lays a solid foundation for the accurate trajectory extraction of subsequent schlieren imaging and the three-dimensional positioning of the leak point, and overall improves the efficiency and accuracy of leak detection of the electrical connection panel of the energy storage battery cover.

[0010] The present invention is further configured as follows: In step S2, an infrared and visible light unified calibration plate is used to complete the hand-eye joint calibration of the infrared imager, the orthogonally arranged schlieren imaging system, and the structured light 3D scanner, and to establish a unified transformation matrix of the pixel coordinate system, the camera coordinate system, and the detection space world coordinate system. Based on the unified transformation matrix, the coordinate mapping relationship between the 3D electronic model and the detection space is established.

[0011] By adopting the above technical solution and using a unified infrared and visible light calibration board, a unified and high-precision calibration benchmark is provided for infrared imagers, orthogonally arranged schlieren imaging systems, and structured light 3D scanners. The hand-eye joint calibration technology can accurately determine the relative positions and orientations of these heterogeneous imaging devices and unify them with the detection space world coordinate system. The unified transformation matrix established in this way for the pixel coordinate system, camera coordinate system, and detection space world coordinate system greatly simplifies the complexity of multimodal data fusion and ensures that two-dimensional image data obtained from different imaging devices can be accurately converted into three-dimensional spatial coordinates. On this basis, the coordinate mapping relationship between the three-dimensional electronic model and the detection space is established based on this unified transformation matrix. This allows the three-dimensional spatial coordinates of the leak point located by the schlieren imaging system in subsequent steps to be accurately mapped onto the three-dimensional electronic model of the panel under test, thereby achieving accurate three-dimensional annotation of the leak point. This method avoids the tedious process of performing complex calibration and coordinate transformation for each imaging device separately, significantly improves calibration efficiency and positioning accuracy, and provides a reliable spatial reference for subsequent leak point analysis and repair.

[0012] The present invention is further configured such that, in step S4, the orthogonally arranged schlieren imaging system is equipped with three sets of schlieren imaging units. The three sets of schlieren imaging units are arranged with the detection space world coordinate system as the reference. The optical axes of the three sets of schlieren imaging units are orthogonal to each other and are all coplanar with the detection plane of the panel under test. Each set of schlieren imaging units is equipped with independent schlieren optical components and image acquisition units.

[0013] By adopting the above technical solution, when performing schlieren imaging detection on suspected leak areas, three sets of schlieren imaging units with pairwise orthogonal optical axes and coplanar with the detection plane of the panel under test can be configured to simultaneously capture images of the leaking jet from three mutually perpendicular perspectives. This multi-view, orthogonal imaging method greatly enriches the spatial information of the leaking jet and effectively avoids the information loss and ambiguity of three-dimensional reconstruction that may be caused by a single or dual perspective. The independent schlieren optical components and image acquisition units of each unit ensure the independent optimization and synchronization of image acquisition from each perspective, thereby improving the accuracy of leaking jet trajectory extraction. Finally, by performing reverse source tracing fitting and spatial intersection calculation on the schlieren images of these three sets of orthogonal perspectives, the three-dimensional spatial coordinates of the leak point can be determined with higher accuracy and robustness, providing more reliable positioning information for subsequent three-dimensional labeling of the leak point and rework operations, and significantly improving the precision level of airtightness detection of the electrical connection panel of the energy storage battery cover.

[0014] The present invention is further configured such that: in step S4, based on the coordinate range of the suspected leakage area, three sets of schlieren imaging units are controlled to perform targeted imaging acquisition of the suspected leakage area, and schlieren image acquisition is not performed in non-suspected areas, and the acquisition field of view of the three sets of schlieren imaging units is synchronously controlled to be completely matched with the coordinate range of the suspected leakage area.

[0015] By adopting the above technical solution, after the suspected leak area is initially identified by infrared thermal imaging, the schlieren imaging system can accurately focus the acquisition resources on this specific area. This targeted imaging acquisition avoids unnecessary image acquisition of non-suspected areas, significantly reduces the amount of data in the schlieren image, thereby reducing the computational burden of data storage and subsequent image processing, and improving processing efficiency. At the same time, by making the acquisition field of view of the schlieren imaging unit completely match the coordinate range of the suspected leak area, it can ensure that higher resolution and finer imaging of the suspected leak area can be achieved with limited imaging resources, thereby capturing the details of the leaking jet more clearly. This provides high-quality raw data for subsequent extraction of the leaking jet trajectory and precise location of the leak point, ultimately improving the positioning accuracy and detection efficiency of the three-dimensional spatial coordinates of the leak point.

[0016] The present invention is further configured as follows: in step S5, a clean foreground image of the leaking jet is extracted from the schlieren image of each orthogonal viewpoint, and then the optical flow vector field of the leaking jet in the foreground image is calculated by optical flow method. Based on the flow direction of the optical flow vector field, the coordinates of the jet source under each viewpoint are obtained by reverse tracing and fitting.

[0017] By adopting the above technical solution, when processing the schlieren images from various orthogonal perspectives, a background modeling algorithm is first used to effectively filter out static background noise and dynamic environmental interference, thereby obtaining a clean foreground image of the leaking jet. This processing step significantly improves the signal-to-noise ratio of the image and avoids interference from non-leakage factors on the extraction of the jet trajectory. On this basis, the optical flow vector field of the leaking jet in the foreground image is accurately calculated using the optical flow method, which can accurately capture the flow direction and velocity information of the leaking gas. Based on these precise flow directions, reverse source tracing fitting is performed, which can accurately determine the two-dimensional source coordinates of the leaking jet from each perspective. Compared with directly extracting the trajectory from the original schlieren image, this solution greatly improves the accuracy and reliability of the leaking jet trajectory extraction through refined image preprocessing and motion analysis. This provides a more accurate input for obtaining the three-dimensional spatial coordinates of the leak point through spatial intersection fitting from various orthogonal perspectives, and ultimately improves the overall accuracy and efficiency of locating the leak point on the electrical connection panel of the energy storage battery cover.

[0018] The present invention is further configured as follows: In step S5, when there are multiple leakage points, the independent jet trajectories of adjacent leakage points are distinguished by cluster analysis of multiple sets of jet trajectories extracted from the orthogonal view schlieren images, and reverse source tracing fitting and spatial intersection calculation are performed on each set of independent jet trajectories to obtain the independent three-dimensional spatial coordinates corresponding to each leakage point.

[0019] By adopting the above technical solution, this application can effectively solve the problem that traditional methods are difficult to accurately distinguish and locate when there are multiple leakage points in the panel under test, especially when there are adjacent leakage points. By introducing cluster analysis, the system can intelligently group the jet trajectories from different leakage points, thereby clearly identifying each group of independent leakage trajectories. On this basis, reverse source tracing fitting and spatial intersection calculation are performed on each group of independent trajectories, ensuring that even in complex multi-leakage scenarios, the independent three-dimensional spatial coordinates of each leakage point can be accurately obtained. This greatly improves the precision and accuracy of leakage detection, and provides more reliable and efficient technical support for the quality control and defect repair of the electrical connection panel of the energy storage battery cover. It avoids misjudgment or missed judgment caused by confusion of multiple leakage points, thereby improving product yield and production efficiency.

[0020] The present invention is further configured as follows: after step S6 is completed, for the panels under test that have completed testing within the same production batch, the leakage point three-dimensional marking results are cumulatively recorded, and the leakage frequency in each preset structural area of ​​the panel under test is counted. When the leakage frequency in any structural area reaches the preset frequency threshold, for the subsequent unqualified parts of the batch that have passed the initial airtightness screening, infrared thermal imaging and / or schlieren imaging detection is preferentially performed on the structural areas that have reached the preset frequency threshold. After the structural areas that have reached the preset frequency threshold have been inspected, infrared thermal imaging detection of the entire area of ​​the panel under test is then performed.

[0021] By adopting the above technical solution, this application can fully utilize historical test data within the same production batch when performing airtightness testing on the electrical connection panel of the energy storage battery cover. By accumulating and statistically analyzing the leakage frequency of each preset structural area through three-dimensional marking of leakage points, the system can intelligently identify high-risk leakage areas within the batch. When the leakage frequency of a specific area reaches a preset threshold, the subsequent inspection process for non-conforming parts will be optimized, prioritizing infrared thermal imaging and / or schlieren imaging detection of these high-risk areas. This targeted detection strategy significantly improves detection efficiency and reduces unnecessary full-area scanning time. Especially when there are common leakage problems in the production batch, it can quickly locate the problem area, thereby accelerating the decision-making process for rework or process adjustment. At the same time, full-area detection is still performed after the priority area inspection is completed, ensuring the comprehensiveness of the inspection and avoiding the omission of potential leaks in other areas, thus achieving a balance between efficiency and accuracy.

[0022] The present invention is further configured as follows: an airtightness detection system for an electrical connection panel of an energy storage battery cover, comprising an airtightness clamping assembly, a pressure detection assembly, and a control module. The airtightness clamping assembly is used to clamp and seal the panel to be tested to form a sealed detection cavity. The pressure detection assembly is connected to the sealed detection cavity and is used to detect the pressure change of the sealed detection cavity to complete the initial screening of overall airtightness. The detection system further comprises a carbon dioxide supply assembly, an imaging assembly, and a processing module. The carbon dioxide supply assembly is connected to the sealed detection cavity and is used to fill the sealed detection cavity with carbon dioxide as a tracer gas. The imaging components include an infrared thermal imaging component and at least two sets of orthogonally distributed schlieren imaging components. The operating band of the infrared thermal imaging component matches the characteristic absorption band of carbon dioxide, and its field of view covers the entire area of ​​the panel under test. It is used to image the entire area of ​​the panel under test and output the suspected leak area. The schlieren imaging component is used to simultaneously acquire schlieren images of the leaking jet from various orthogonal viewpoints of the suspected leak area. The processing module processes the infrared imaging data and the schlieren imaging data to extract the leaking jet trajectory. Based on the multi-view leaking jet trajectory fitting, the three-dimensional spatial coordinates of the leak point are obtained. The three-dimensional spatial coordinates of the leak point are then mapped onto the three-dimensional electronic model of the panel under test to complete the stereoscopic annotation of the leak point.

[0023] By adopting the above technical solution, combining carbon dioxide tracer gas with infrared thermal imaging technology that matches characteristic absorption bands, and introducing three sets of orthogonally distributed schlieren imaging systems and a linked windproof shield assembly, the system achieves precise location and three-dimensional spatial coordinate fitting of leak points in metal panels. Since carbon dioxide, as a low-cost tracer gas, avoids the high cost problem of helium, and the infrared thermal imaging and characteristic absorption band matching effectively overcome the interference of metal surface reflection, the three sets of orthogonally distributed schlieren imaging systems achieve precise extraction of the leak jet trajectory through multi-view synchronous acquisition, and the linked windproof shield assembly isolates environmental airflow interference. Therefore, the system can quickly and accurately locate leak points, meet the needs of large-scale full inspection on production lines, and significantly improve detection efficiency and reliability.

[0024] The present invention is further configured such that the arithmetic processing module includes: The synchronization control unit is used to synchronize the gas supply action with the imaging acquisition. The image processing unit is used to perform noise reduction and feature extraction on infrared imaging data and schlieren imaging data, and output the leakage jet trajectory. A three-dimensional positioning unit is used to obtain the three-dimensional spatial coordinates of the leak point based on multi-view leak jet trajectory fitting. The annotation generation unit is used to complete the three-dimensional annotation of the leak points.

[0025] By adopting the above technical solution, the computing module is finely divided into a synchronous control unit, an image processing unit, a 3D positioning unit, and a labeling generation unit, thereby significantly improving the intelligence and automation level of the entire airtightness detection system. The synchronous control unit ensures precise timing coordination between gas supply and multimodal imaging acquisition, avoiding misalignment and invalidity in data acquisition and guaranteeing the accuracy of subsequent data processing. The image processing unit can efficiently reduce noise and extract features from infrared and schlieren images, especially accurately identifying the leakage jet trajectory in complex backgrounds, providing high-quality input for subsequent precise positioning. The 3D positioning unit, based on multi-view trajectory information, achieves high-precision 3D spatial positioning of the leak point through advanced geometric algorithms, solving the problem that traditional methods cannot accurately obtain the spatial location of the leak point. Finally, the labeling generation unit intuitively maps the positioning results onto the 3D electronic model and generates detailed detection reports and traceability data, greatly improving the readability, traceability, and efficiency of guiding production rework of the detection results. This modular design makes the system more efficient and stable in processing complex multi-source data and can output high-precision, high-value detection results, thereby effectively improving the quality control level of the electrical connection panel on the energy storage battery cover.

[0026] The present invention has significant technical effects due to the adoption of the above technical solutions: The airtightness detection method and system for the electrical connection panel of the energy storage battery cover provided in this application uses carbon dioxide tracer gas, infrared thermal imaging and schlieren imaging technology after the initial inspection, combined with coordinate mapping to accurately locate the leak point. It has the advantages of being able to accurately locate the leak point, improving detection efficiency and product quality. Attached Figure Description

[0027] Figure 1 This is a step-by-step diagram of the method for testing the airtightness of the electrical connection panel on the top cover of an energy storage battery; Figure 2 This is a schematic diagram of the airtight clamping assembly structure; Figure 3 This is a partial cross-sectional view of the airtight clamping assembly.

[0028] The parts referred to by the numbers in the above attached diagrams are as follows: 1. Upper sealing plate; 2. Lower sealing plate; 3. Control cylinder; 4. Sealing ring; 5. Upper cover electrical connection panel; 6. Vent hole. Detailed Implementation

[0029] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0030] Example: Traditional methods for detecting the airtightness of electrical connection panels on the top cover of energy storage batteries, such as differential pressure method, helium mass spectrometry and traditional thermal imaging method, have limitations when facing the need to detect ultra-micro leaks in metal panels. These limitations include the inability to accurately locate the leak point, low detection efficiency, high cost, and susceptibility to false alarms caused by surface reflection and metal properties. These limitations make it difficult to meet the needs of large-scale full inspection on production lines.

[0031] This application proposes a method for airtightness testing of the electrical connection panel of an energy storage battery cover. This method achieves precise three-dimensional positioning and visual annotation of the leak point through multi-stage and multi-modal detection. Specifically, the method includes clamping and sealing the panel under test and completing an overall initial airtightness inspection. Subsequently, a three-dimensional electronic model is preloaded and the coordinates of the imaging system and the detection space are calibrated. For unqualified parts determined in the initial inspection, carbon dioxide gas is injected into the sealed detection chamber as a tracer gas. Infrared thermal imaging matching the characteristic absorption band of carbon dioxide is used to image the entire area of ​​the panel under test to extract the suspected leak area. For the suspected leak area, the orthogonally arranged schlieren imaging system is controlled to simultaneously acquire schlieren images of the leak jet from each orthogonal viewpoint. Then, the leak jet trajectory is extracted from the schlieren images of each orthogonal viewpoint. The jet source coordinates of the leak jet trajectory at each viewpoint are obtained by reverse source tracing fitting. Then, the three-dimensional spatial coordinates of the leak point are obtained by spatial intersection fitting. Finally, the three-dimensional spatial coordinates of the leak point are mapped into the three-dimensional electronic model to complete the three-dimensional annotation of the leak point.

[0032] The process involves clamping and sealing the panel under test to form a sealed testing chamber, then filling the chamber with gas. A preliminary airtightness test is performed by detecting pressure changes within the chamber, thus distinguishing between qualified and unqualified parts. Specifically, the panel under test can be placed in an airtight fixture, and its edges are pressed together mechanically or pneumatically to form a sealed testing chamber isolated from the external environment. Subsequently, a testing gas, such as air or nitrogen, is filled into the chamber through a gas supply device. A high-precision pressure sensor is used to monitor pressure changes inside the chamber in real time. If the pressure drop exceeds a preset threshold within a preset testing time, the panel under test is determined to be unqualified; otherwise, it is determined to be qualified.

[0033] This method preloads a 3D electronic model corresponding to the panel under test and completes coordinate calibration of at least two sets of orthogonally arranged schlieren imaging systems and the detection space to establish a unified coordinate mapping relationship between the 3D electronic model, the schlieren imaging system, and the detection space. For example, the digital 3D model of the panel under test can be pre-stored in the database of the detection system and automatically loaded according to the model of the panel under test. In order to achieve accurate spatial positioning, the schlieren imaging system needs to be calibrated. This can be done by placing a calibration target with known geometric features in the detection space and using the schlieren imaging system to image it from multiple angles. The feature points on the calibration target are identified by the image processing algorithm, and the transformation matrix between the schlieren imaging system and the detection space is calculated. At the same time, by placing a physical reference object that matches the 3D electronic model in the detection space and performing image recognition and matching, the mapping relationship between the 3D electronic model and the detection space can be established.

[0034] Carbon dioxide gas is introduced into the sealed detection chamber as a tracer gas. For non-conforming parts determined by the initial airtightness inspection, infrared thermal imaging matching the characteristic absorption band of carbon dioxide is used to image the entire area of ​​the panel under test and extract suspected leakage areas. Specifically, for the panel under test that is determined to be non-conforming in the initial airtightness inspection, the original gas in its sealed detection chamber can be discharged and then pure carbon dioxide gas can be refilled, or carbon dioxide gas can be added on the basis of the original gas to achieve a sufficient tracer concentration. An infrared thermal imager is configured to operate in a wavelength range matching the characteristic absorption band of carbon dioxide, such as 4.2 micrometers to 4.4 micrometers. The infrared thermal imager scans and images the entire surface of the panel under test. By analyzing the local temperature or radiation intensity anomalies caused by carbon dioxide leakage in the infrared image, the image processing software can automatically identify and delineate potential leakage areas and mark them as suspected leakage areas.

[0035] For the suspected leak area, the orthogonally deployed schlieren imaging system is controlled to simultaneously acquire schlieren images of the leaking jet from various orthogonal viewpoints. For example, at least two sets of schlieren imaging systems are installed in the detection space with their optical axes orthogonal to each other, such as along the X-axis, Y-axis, or Z-axis. After the infrared thermal imaging identifies the coordinate range of the suspected leak area, the control system will drive the schlieren imaging system to perform targeted positioning based on the coordinate information, so that its field of view can accurately cover the suspected leak area. Subsequently, the schlieren imaging systems at various orthogonal viewpoints are triggered synchronously to ensure that schlieren images of the leaking jet from different viewpoints are captured at the same time.

[0036] The leakage jet trajectory is extracted from the schlieren images of each orthogonal viewpoint. The jet source coordinates of the leakage jet trajectory under each viewpoint are obtained by reverse source tracing fitting. The three-dimensional spatial coordinates of the leakage point are obtained by spatial intersection fitting of each orthogonal viewpoint. Specifically, for the schlieren images acquired from each orthogonal viewpoint, image preprocessing is first performed, such as noise reduction and enhancement. Then, the outline and flow direction of the leakage jet are accurately extracted from the image using image segmentation algorithms or optical flow methods, thereby obtaining the leakage jet trajectory. Based on these trajectories, reverse source tracing fitting is performed using a fluid dynamics model or particle tracking algorithm to calculate the starting point coordinates of the jet under each viewpoint, i.e., the jet source coordinates. Finally, these jet source coordinates from different orthogonal viewpoints are input into the three-dimensional reconstruction algorithm. Spatial intersection fitting is performed using triangulation or multi-view geometry principles to accurately calculate the coordinates of the leakage point in three-dimensional space.

[0037] Finally, the three-dimensional spatial coordinates of the leak point are mapped onto the aforementioned three-dimensional electronic model to complete the three-dimensional annotation of the leak point. For example, the three-dimensional coordinates of the leak point obtained through spatial intersection fitting are transmitted to the three-dimensional electronic model processing system. This system uses a pre-established unified coordinate mapping relationship to accurately locate the leak point on the three-dimensional electronic model of the panel under test. On the model, the leak point can be visualized as a three-dimensional mark with a specific color, size, or shape, such as a red sphere or arrow, thereby achieving intuitive and accurate three-dimensional annotation of the leak point.

[0038] The above proposes a method for performing a preliminary airtightness inspection on the panel under test and for subsequent fine-tuning leak location on non-conforming parts identified in the preliminary inspection. However, in the implementation process, if the gas used in the preliminary inspection stage is different from the tracer gas used in the subsequent fine-tuning inspection, additional gas switching or purging steps may be required, which increases the complexity and time cost of the inspection process. At the same time, if the gas pressure in the fine-tuning inspection stage is the same as that in the preliminary inspection stage, it may not be able to effectively enhance the leak signal, affecting the ability of subsequent infrared thermal imaging to capture and locate minute leaks.

[0039] In step S1, carbon dioxide gas is introduced into the gas filling the sealed testing chamber for the initial airtightness inspection. Step S1 is the initial inspection stage of the airtightness inspection, and its purpose is to quickly screen out unqualified parts with leaks. Using carbon dioxide gas as the filling gas in the initial inspection stage means that the same gas is used as the detection medium throughout the entire inspection process, from the initial inspection to the subsequent refined leak location. Carbon dioxide gas is non-toxic, non-corrosive, easy to obtain, and relatively low in cost. At the same time, its molecular weight is moderate, and it can form a jet that can be detected by a specific infrared band when leaking. This unified gas usage strategy simplifies the inspection process and avoids the system complexity or cross-contamination problems that may be caused by switching between different gases.

[0040] In step S3, for the defective parts determined by the initial airtightness inspection, the gas pressure injected into the sealed detection chamber is higher than that injected in step S1. Step S3 is the first step in the refined leak location for the defective parts, that is, using infrared thermal imaging to extract the suspected leak area. In this stage, the gas pressure injected into the sealed detection chamber is higher than that in step S1, which aims to increase the pressure difference inside and outside the leak point. The increase in pressure difference will lead to an increase in the velocity and flow rate of the leak jet, making the leak signal, such as carbon dioxide gas, which absorbs the signal in the infrared band more significant, thereby improving the detection sensitivity of infrared thermal imaging for small leaks. The higher pressure difference helps to form a clearer and easier-to-capture leak jet, providing a better foundation for the subsequent accurate trajectory extraction and source location of schlieren imaging.

[0041] This application further proposes that, when establishing coordinate mapping relationships, an infrared and visible light unified calibration plate is used to complete the hand-eye joint calibration of the infrared imager, the orthogonally arranged schlieren imaging system, and the structured light 3D scanner, and to establish a unified transformation matrix of the pixel coordinate system, camera coordinate system, and detection space world coordinate system. Based on the unified transformation matrix, the coordinate mapping relationship between the 3D electronic model and the detection space is established.

[0042] Specifically, the infrared-visible light unified calibration board is a specially designed calibration tool whose surface pattern or material can be clearly imaged under both visible and infrared spectra. For example, checkerboard patterns and circular arrays can be made of materials with specific geometric patterns and good reflection or absorption characteristics for both infrared and visible light. This calibration board provides a unified reference benchmark for imaging devices in different operating bands, ensuring that all devices can be calibrated based on the same physical reference in multimodal imaging scenarios.

[0043] By employing hand-eye joint calibration technology, the precise spatial position and orientation of three types of heterogeneous imaging devices—infrared imagers, orthogonally arranged schlieren imaging systems, and structured light 3D scanners—relative to the detection space world coordinate system can be accurately determined. This calibration process typically involves placing an infrared-visible light system calibration board at multiple different positions and orientations within the detection space, and simultaneously acquiring image or point cloud data of each imaging device relative to the calibration board. By processing and optimizing these multimodal data, the internal parameters of each imaging device, such as focal length, principal point, and distortion coefficient, as well as their external parameters relative to the detection space world coordinate system, can be obtained. Based on this, a unified transformation matrix is ​​established for the pixel coordinate system, camera coordinate system, and detection space world coordinate system. This unified transformation matrix integrates all transformation relationships from image pixel coordinates to the 3D coordinates of each camera itself, and then to the detection space world coordinate system. For example, for each camera, the pixel coordinates are first converted to the 3D ray direction in the camera coordinate system using its intrinsic parameters, and then the 3D points in the camera coordinate system are converted to the detection space world coordinate system using its extrinsic parameters.

[0044] Based on a unified transformation matrix, the coordinate mapping relationship between the three-dimensional electronic model and the detection space is established. This is usually done by identifying the positions of preset feature points on the panel under test in the world coordinate system of the detection space and matching them with the corresponding feature points in the three-dimensional electronic model, thereby calculating the precise transformation relationship between the three-dimensional electronic model and the world coordinate system of the detection space.

[0045] This application further proposes that in step S4, the orthogonally arranged schlieren imaging system is configured with three sets of schlieren imaging units. These three sets of schlieren imaging units are arranged with the detection space world coordinate system as the reference. The optical axes of the three sets of schlieren imaging units are orthogonal to each other and coplanar with the detection plane of the panel under test. Each set of schlieren imaging units is equipped with an independent schlieren optical component and image acquisition unit. The schlieren imaging system is used to visualize the density gradient changes caused by gas leakage, thereby capturing the trajectory of the leaking jet. With three sets of schlieren imaging units, compared to only two sets, image information of the leaking jet can be obtained from more dimensions. The three sets of units work together to provide a more comprehensive spatial perspective, effectively avoiding blind spots or information loss that may exist due to a single or dual perspective. This provides richer and more reliable raw data for subsequent leaking jet trajectory extraction and three-dimensional spatial coordinate fitting. The detection space world coordinate system is the unified reference system for the entire detection environment; the position information of all sensors and the panel under test will be mapped to this coordinate system. Three sets of schlieren imaging units are arranged based on this coordinate system, ensuring that the relative positional relationship between each imaging unit is accurately known, and a unified coordinate mapping relationship is established with the three-dimensional electronic model of the panel under test. This unified coordinate reference is the basis for achieving high-precision three-dimensional positioning and stereo annotation. It ensures that image data acquired from different perspectives can be accurately correlated and fused. The pairwise orthogonal optical axes mean that the central optical axes of the three sets of schlieren imaging units are perpendicular to each other in space, and can be arranged along the three orthogonal directions of X, Y, and Z respectively. This orthogonal layout can capture the different projections of the leaking jet in three-dimensional space to the greatest extent, thereby providing the most comprehensive spatial information. At the same time, the fact that each unit is coplanar with the detection plane of the panel under test means that the field of view center or key imaging area of ​​each imaging unit is at the same height or in the same reference plane as the leak detection plane of the panel under test. This helps to ensure that the leaking jet is clearly and completely imaged in the key detection area and simplifies the subsequent image processing and three-dimensional reconstruction algorithms.

[0046] Independent schlieren optics typically include core components such as a light source, blade, and concave mirror, used to achieve the schlieren effect and form a clear image of the leaking jet. An independent image acquisition unit is responsible for converting the image formed by the optical components into a digital signal. This independent configuration allows each schlieren imaging unit to be independently optimized and adjusted according to its specific viewing angle and imaging requirements, such as exposure time, gain, and focal length. In addition, the independent image acquisition unit can achieve synchronous acquisition of images from different viewing angles, ensuring that different sides of the leaking jet are captured at the same time point, which is crucial for the accurate three-dimensional reconstruction of dynamic leaking jets.

[0047] In step S4, based on the coordinate range of the suspected leak area, three sets of schlieren imaging units are controlled to perform targeted imaging acquisition of the suspected leak area. Schlieren image acquisition is not performed on non-suspected areas. Simultaneously, the acquisition field of view of the three sets of schlieren imaging units is controlled to perfectly match the coordinate range of the suspected leak area. Specifically, the coordinate range of the suspected leak area is the area with high carbon dioxide concentration or temperature anomalies identified and extracted by image processing algorithms after infrared thermal imaging of the entire panel under test. This coordinate range can be a pixel area in a two-dimensional image coordinate system or a three-dimensional spatial coordinate range after coordinate mapping. Targeted imaging acquisition refers to precisely adjusting the imaging parameters and / or physical position of the schlieren imaging units according to the determined coordinate range of the suspected leak area, so that their acquisition field of view concentrates on covering this specific area. This can be achieved by adjusting the focal length and zoom ratio of the schlieren imaging units, or by moving the imaging units through actuators such as robotic arms. Meanwhile, schlieren image acquisition is not performed on non-suspected areas. This means that the schlieren imaging system will skip or not record image data of areas that have not been identified as suspected leaks by infrared thermal imaging. This can be achieved by setting area filtering conditions in the image acquisition control logic, or by avoiding imaging of non-suspected areas through physical occlusion, adjusting the scanning path of the imaging unit, etc. In addition, the system also synchronously controls the acquisition field of view of the three sets of schlieren imaging units to be completely matched with the coordinate range of the suspected leak area. Synchronous control means that at the same time point or within a very short time interval, the actions of the three sets of schlieren imaging units are coordinated so that their acquisition field of view is simultaneously and accurately aligned with the suspected leak area. Complete matching means that the effective acquisition area of ​​the schlieren imaging unit coincides with the boundary of the suspected leak area as much as possible to ensure that all potential leak information is captured, while avoiding the acquisition of data from irrelevant areas. This usually requires a high-precision motion control system and a real-time image feedback mechanism for calibration and adjustment.

[0048] In response, this application further proposes that in step S5, a background modeling algorithm is used to filter out static background noise and environmental dynamic interference from the schlieren images of each orthogonal viewpoint, and a clean foreground image of the leaking jet is extracted. Then, the optical flow vector field of the leaking jet in the foreground image is calculated by optical flow method, and the coordinates of the jet source under each viewpoint are obtained by reverse tracing and fitting based on the flow direction of the optical flow vector field.

[0049] Specifically, background modeling algorithms are image processing techniques that classify pixels in an image sequence into background or foreground. By learning and building a background model of the scene, these algorithms can effectively identify and separate moving or changing objects (foreground) from the relatively stable background. Common background modeling methods include Gaussian mixture models, frame difference methods, and median filtering. By applying background modeling algorithms, leaking jets can be accurately identified from complex fingerprint images, providing a clean data source for subsequent analysis. Static background noise is typically a fixed interference within the fingerprint imaging system itself or the detection environment, such as light. Dust, scratches, and shadows from mounting brackets on the components are removed, while environmental dynamic interference refers to non-leakage-related air disturbances, minor vibrations, or changes in lighting that may occur during the detection process. Through background modeling algorithms, these interferences are identified and removed from the original schlieren image, ensuring that subsequent processing targets only the actual leaking jet information. This significantly improves image purity. After background modeling and filtering out various interferences, the image retains only the area formed by light refraction or absorption changes caused by the leaking jet—a pure foreground image of the leaking jet. This foreground image clearly outlines the shape, location, and motion characteristics of the leaking jet, providing a basis for subsequent processing. The optical flow method provides high-quality input data for optical flow calculation and reverse source tracing. Based on this, optical flow is a technique for estimating pixel motion in image sequences. It calculates the motion vector of each pixel by analyzing the spatiotemporal changes in pixel intensity, forming an optical flow vector field. Each vector represents the displacement direction and velocity of the corresponding pixel between consecutive frames. In the schlieren image of a leaking jet, the optical flow method can accurately capture the direction and velocity information of the gas flow inside the jet, providing crucial kinematic data for reverse inference of the jet source. After obtaining a clean foreground image of the leaking jet, the optical flow method is applied to process the foreground image sequence, and the motion vector is calculated using optical flow. By calculation, a two-dimensional vector field can be obtained, where each vector represents the instantaneous motion direction and velocity of the gas particles in the leaking jet. This optical flow vector field intuitively shows the dynamic characteristics of the leaking jet and is the basis for accurate reverse tracing. After obtaining the optical flow vector field of the leaking jet, the flow direction information of these vector fields is used to perform reverse deduction. Specifically, it is possible to trace along the opposite direction of each optical flow vector and fit these reverse tracing paths. Theoretically, all reverse tracing paths should converge at the source of the leak. Through this reverse tracing fitting, the two-dimensional source coordinates of the leaking jet at each orthogonal perspective can be accurately determined.

[0050] In step S5 above, when there are multiple leakage points, cluster analysis is performed on multiple sets of jet trajectories extracted from each orthogonal viewpoint schlieren image to distinguish the independent jet trajectories of adjacent leakage points. For each set of independent jet trajectories, reverse source tracing fitting and spatial intersection calculation are performed to obtain the independent three-dimensional spatial coordinates corresponding to each leakage point.

[0051] Specifically, when two or more gas leak points exist simultaneously on the panel under test, each leak point will generate an independent leak jet. These jets appear as different flow trajectories in the schlieren image. In order to accurately identify and distinguish these trajectories, this application adopts clustering analysis technology. Clustering analysis is a data mining method used to group objects in a dataset so that objects within the same group are similar to each other, while objects in different groups are quite different. Here, clustering analysis is applied to all leak jet trajectories extracted from schlieren images from different orthogonal viewpoints. Its purpose is to classify trajectories originating from the same leak point into one category based on the spatial location, direction, and morphology of the jet trajectory, thereby distinguishing the trajectories of different leak points. Clustering algorithms such as K-means and DBSCAN can be used to measure and group similarity based on the geometric center, direction vector, or pixel distribution characteristics of the jet trajectory.

[0052] Cluster analysis can clearly identify the specific leak source corresponding to each set of trajectories, thus distinguishing the independent jet trajectories of adjacent leak points. While adjacent leak points may exhibit some overlap or proximity in their jet trajectories on the image, they can still be separated through sophisticated clustering algorithms and feature analysis. For example, a clustering distance threshold can be set, or the dynamic characteristics of local consistency in jet vector fields, such as optical flow vector fields, can be combined to enhance the distinguishing ability, ensuring that each leak point has its own unique set of trajectories. Once multiple independent sets of jet trajectories are distinguished through cluster analysis, each set of trajectories will be executed independently. The reverse source fitting and spatial intersection calculation are performed by using the reverse source fitting to deduce the starting point of the jet, i.e., the coordinates of the jet source, based on the flow direction of the jet. The spatial intersection calculation uses the coordinates of the jet source from different perspectives provided by at least two sets of orthogonally arranged schlieren imaging systems to accurately calculate the coordinates of the leak point in three-dimensional space through the principle of geometric triangulation. After the above reverse source fitting and spatial intersection calculation for each set of independent jet trajectories, multiple different three-dimensional spatial coordinates will be obtained. Each coordinate corresponds precisely to an independent leak point on the panel under test.

[0053] This application further proposes a method to optimize the detection process. Specifically, after step S6 is completed, that is, after the leakage points have been stereo-annotated in the three-dimensional electronic model, the system will accumulate and record the three-dimensional annotation results of the leakage points of the panels under test that have been tested in the same production batch. This means that the three-dimensional spatial coordinates of the leakage points of each panel under test and its related attribute information such as leakage size and type will be systematically collected and stored to form a historical leakage database within the batch. On this basis, the system will count the frequency of leakage in each preset structural area of ​​the panel under test. These preset structural areas can be specific welding areas, sealing strip areas, connector interface areas, etc. on the panel. They are pre-divided according to the design and manufacturing process of the panel. By mapping the accumulated leakage points to these preset areas, the system can calculate the number of times each area leaks in the current batch.

[0054] When the frequency of leakage in any structural area reaches a preset frequency threshold, that area is identified as a high-risk leakage area in that batch. This preset frequency threshold is a configurable parameter that can be set according to product quality requirements, historical production data, or process experience. Once an area reaches this threshold, the testing process will be optimized for subsequent non-conforming parts in that batch that have passed the initial airtightness screening: the system will prioritize infrared thermal imaging and / or schlieren imaging detection on structural areas that reach the preset frequency threshold. This means that the field of view or scanning path of the imaging component will first focus on these high-risk areas for localized, targeted, and detailed detection. After the structural areas that reach the preset frequency threshold have been inspected, if leakage is required or not found, the system will then perform infrared thermal imaging detection on the entire area of ​​the panel under test to ensure comprehensive detection.

[0055] In step S6 above, when completing the three-dimensional annotation of the leak point in the 3D electronic model, the location information and attribute information of the leak point are simultaneously annotated. A test report and rework instructions with three-dimensional annotations are generated simultaneously. At the same time, the pressure data, imaging data, leak point coordinates, and annotation results of the entire testing process are simultaneously uploaded to the production line manufacturing execution system to complete the traceability of the quality data throughout the entire lifecycle of the panel under test. The location information of the leak point may include, but is not limited to, a specific area on the panel under test, such as a weld, sealing ring, or connector interface, or its relative position to a specific reference feature point on the panel under test. This location information can be automatically identified and associated through preset area division rules or based on the geometric features of the 3D electronic model. The attribute information of the leak point can cover the type of leak, such as microcracks, holes, or assembly gaps, the size of the leak, the environmental conditions when the leak occurred, the corresponding production batch, production station, and operator, etc. This attribute information can be obtained through manual input, integration with other sensor data, or automatic acquisition through algorithm analysis.

[0056] Based on this, a test report with 3D annotations and a rework instruction manual are generated simultaneously. The test report typically includes the test time, product batch number, pass or fail test results, 3D spatial coordinates of the leak point, detailed location information, attribute information, and relevant infrared thermal imaging and schlieren imaging data. This report can be output in a structured data format or a highly readable document format. The rework instruction manual provides specific rework steps, required tools, rework standards, and precautions for the detected leak point. For example, if the leak point is located at a weld point, the instruction manual may detail the welding process parameters and inspection standards; if it is located in a sealing structure, it may indicate the replacement of the seal or reassembly. The instruction manual may include illustrations, 3D model views, or video links to intuitively guide operators in rework.

[0057] The pressure data, imaging data, leak point coordinates, and annotation results from the entire detection process are synchronously uploaded to the manufacturing execution system (MES) on the production line. The detection process data includes, but is not limited to, the pressure change curve of the sealed detection chamber in step S1, the raw infrared thermal imaging image and processed suspected leak area data in step S3, the raw schlieren image data from each orthogonal viewing angle in step S4, the leak jet trajectory extracted in step S5, the jet source coordinates, and the finally determined three-dimensional spatial coordinates of the leak point. The annotation results refer to the location information, attribute information, and stereoscopic annotation status of the leak point in the three-dimensional electronic model. Through a standardized data interface, automatic data transmission and sharing between the detection system and the MES system are achieved, ensuring the real-time performance and accuracy of the data, and synchronously annotating the leak point. The system generates location and attribute information, along with 3D-annotated inspection reports and rework instructions. This provides clear and specific guidance for subsequent rework and repair, significantly improving rework efficiency and accuracy, and reducing errors from human judgment. Simultaneously, it uploads pressure data, imaging data, leak point coordinates, and annotation results from the entire inspection process to the production line's manufacturing execution system, achieving deep integration of inspection data with the production process. This constructs a full lifecycle quality data chain for the panel under test, from production to use, greatly enhancing product quality traceability. This not only helps to quickly locate and resolve quality issues in production but also provides comprehensive and reliable data support for production process optimization and quality management decisions, thereby effectively improving the overall product quality level and production efficiency.

[0058] After step S6 is completed, based on the leak point coordinates, attribute information, and detection data of the batch of panels under test, cluster analysis is used to obtain the structural regions where leaks are concentrated. Combined with the production process parameters corresponding to the concentrated leak regions, process optimization parameters for the panel production process are output, forming a closed-loop process of detection, analysis, and optimization. The leak point coordinates refer to the precise location information of the leak point in three-dimensional space obtained in step S6, usually represented by three-dimensional coordinates. These coordinates are obtained after acquisition by the schlieren imaging system, jet trajectory extraction, reverse source tracing fitting, and spatial intersection calculation, which can accurately reflect the physical location of the leak. Attribute information refers to non-geometric information related to the leak point, such as the type of leak point, the amount of leakage, the time of the leak, the detection batch, and the specific area of ​​the panel. This attribute information can be manually input, preset rules, or combined with other methods. He supplements the data with sensor data. The detection data refers to all relevant data generated during the entire airtightness testing process, including but not limited to pressure change data, infrared thermal imaging data, schlieren image data, jet trajectory data, and environmental parameters. The batch of panels to be tested refers to multiple energy storage battery cover electrical connection panels that have undergone the same testing process within a certain period of time or a production batch. Through cluster analysis, the coordinates and / or attribute information of the leakage points of the batch of panels to be tested are grouped to identify specific physical areas or structural features that repeatedly leak on the panels, i.e., structural areas where leakage is concentrated. For example, if the leakage points of multiple panels are concentrated in a specific weld area, near a connector interface, or on the edge of an injection molded part, these areas will be identified as structural areas where leakage is concentrated. This helps to discover systemic defects in the production process from a macroscopic level.

[0059] The information on concentrated leakage areas obtained through cluster analysis is correlated and compared with the specific production process parameters corresponding to these areas during the production process. Production process parameters refer to process control variables that are directly or indirectly related to the concentrated leakage areas during the production of the electrical connection panel of the energy storage battery cover, such as welding current, welding time, injection temperature, holding time, tightening torque, and amount of sealant applied. This combination usually requires a complete manufacturing execution system or data management system that can trace the specific process parameter records of each panel or each area during production.

[0060] Based on the analysis of concentrated leakage areas and their associated production process parameters, specific suggested values ​​or adjustment directions are proposed for adjusting and improving the production process. This involves outputting process optimization parameters for the panel production stage. For example, if frequent leakage is found in a weld area and is related to insufficient welding current, an optimization suggestion to "increase the welding current" might be output. These optimization parameters can be output through generating optimization reports, directly updating the process parameter database in the manufacturing execution system, or issuing adjustment instructions to production line operators. Using the above technical solution, after accurately locating and three-dimensionally labeling the leakage points of a single non-conforming part, further in-depth mining and analysis of batch production data is conducted. By clustering the leakage point coordinates, attribute information, and detection data of a batch of panels under test, a systematic identification of leakage points in the product can be achieved. The recurring leakage in specific structural areas reveals common defects in the production process. By correlating these concentrated leakage areas with corresponding production process parameters, the specific process steps and parameter deviations causing the leakage can be precisely located. This feedback mechanism based on big data analysis and process parameter correlation transforms the detection results from merely a basis for quality judgment into effective information guiding production process improvement. This allows for the output of process optimization parameters for panel production, effectively solving recurring leakage problems in mass production and preventing the continued occurrence of similar defects. Furthermore, by forming a closed-loop process encompassing detection, analysis, and optimization, it achieves continuous improvement in product quality and optimization of production efficiency, significantly reducing scrap rates and rework costs, and enhancing overall manufacturing capabilities.

[0061] This application proposes an airtightness detection system for the electrical connection panel of an energy storage battery cover. This system utilizes multimodal imaging and three-dimensional spatial positioning technology to achieve precise three-dimensional coordinate positioning and stereoscopic marking of leak points. Specifically, the system includes an airtight clamping component, a pressure detection component, and a control module. The airtight clamping component is configured to clamp and seal the panel under test to form a sealed detection chamber. The pressure detection component is connected to the sealed detection chamber and is configured to detect pressure changes within the sealed detection chamber to complete an initial screening of overall airtightness. The detection system also includes a carbon dioxide supply component, an imaging component, and a processing module. The carbon dioxide supply component is connected to the sealed detection chamber and is configured to fill the chamber with carbon dioxide as a tracer gas. During implementation, the carbon dioxide supply component may include a gas storage tank, a pressure reducing valve, and a flow control valve to precisely control the gas pressure and flow rate, ensuring that the tracer gas concentration meets the detection requirements. The imaging component includes infrared thermal imaging... The imaging components include at least two orthogonally distributed schlieren imaging components. These three sets of schlieren imaging components are arranged along the horizontal X-axis, vertical Y-axis, and depth Z-axis of a preset detection spatial coordinate system. The optical axes of the three sets of schlieren imaging components are perpendicular to each other and intersect at the center of the detection plane of the panel under test. The operating band of the infrared thermal imaging component matches the characteristic absorption band of carbon dioxide, for example, the 4.2-4.4 micrometer band. Its acquisition field of view covers the entire area of ​​the panel under test and is configured to image the entire area of ​​the panel under test and output the suspected leak area. The schlieren imaging components are configured to simultaneously acquire schlieren images of the leaking jet from various orthogonal viewing angles in the suspected leak area. In specific implementation, the infrared thermal imaging component uses a cooled infrared focal plane array detector to improve detection sensitivity and effectively avoid interference from metal surface reflections. The schlieren imaging component includes a light source, a blade, and a high-speed camera, and is configured to capture light refraction caused by changes in gas density, ensuring clear imaging of the leaking jet trajectory.

[0062] See Figure 1 and Figure 2 The airtight clamping assembly includes an upper sealing plate, a lower sealing plate, and a control cylinder for driving the upper sealing plate to move toward the lower sealing plate. An electrical connection panel for the upper cover is disposed between the upper and lower sealing plates. Both the upper and lower sealing plates are provided with sealing rings. The sealing ring of the lower sealing plate abuts against the opening of the electrical connection panel for the upper cover, forming a sealed detection chamber. The lower sealing plate is provided with a vent hole that communicates with the sealed detection chamber. The vent hole is connected to a carbon dioxide supply assembly for filling the sealed detection chamber with detection gas.

[0063] The processing module processes infrared and schlieren imaging data to extract the leakage jet trajectory. It is configured to obtain the three-dimensional spatial coordinates of the leakage point based on multi-view leakage jet trajectory fitting. The three-dimensional spatial coordinates of the leakage point are then mapped onto the three-dimensional electronic model of the panel under test to complete the stereoscopic annotation of the leakage point. The processing module includes an image processing unit for noise reduction and enhancement of the original image; a three-dimensional positioning unit for calculating the leakage point coordinates through a spatial intersection algorithm; and an annotation generation unit for mapping the coordinates onto the three-dimensional model to achieve visual marking of the leakage point. The airtight clamping component is equipped with a linkage windproof cover component. The linkage windproof cover component is linked to the clamping action of the airtight clamping component and is configured to close the detection space of the panel under test during schlieren image acquisition, isolating the leakage jet from the interference of the ambient airflow. The linkage windproof cover component is signal-connected to the main control module. For example, when the airtight clamping component completes the clamping action, the linkage mechanism automatically triggers the windproof cover to close, forming a closed detection environment to ensure that the schlieren imaging is not affected by external airflow disturbances.

[0064] The processing module includes a synchronization control unit, an image processing unit, a 3D positioning unit, and a label generation unit. The synchronization control unit is responsible for coordinating the timing of operations of different hardware components throughout the detection process. Specifically, it ensures that the carbon dioxide supply action and the image acquisition actions of the infrared thermal imaging component and the schlieren imaging component are precisely and synchronously triggered. This can be achieved by sending a unified trigger signal to each component or by controlling a preset time sequence. For example, after the carbon dioxide tracer gas reaches a stable state, the synchronization control unit can immediately trigger the infrared thermal imaging component to perform a full-area scan, and after identifying a suspected leak area, it can precisely trigger the schlieren imaging component to perform local high-precision acquisition. Its function is to ensure the effectiveness and consistency of data acquisition and avoid invalid data or processing difficulties caused by timing errors.

[0065] The image processing unit is mainly used to preprocess and extract features from the raw image data acquired from the imaging component. For infrared imaging data, it may include noise reduction processing such as hotspot enhancement and background subtraction to more clearly identify the thermal anomaly area caused by carbon dioxide leakage. For schlieren imaging data, it may involve image enhancement, edge detection, background modeling algorithms to filter out static background noise and dynamic environmental interference, and optical flow method to calculate the optical flow vector field of the leaking jet, thereby accurately extracting the trajectory information of the leaking jet. The implementation of this unit can be based on a high-performance image processor or a dedicated image processing chip and run the corresponding image processing algorithm library.

[0066] The function of the three-dimensional positioning unit is to use the multi-view leakage jet trajectory information extracted by the image processing unit to accurately calculate the coordinates of the leakage point in three-dimensional space through geometric principles and algorithms. In specific implementation, techniques such as multi-view geometry, triangulation, or spatial intersection fitting can be used. For example, the coordinates of the jet source of the leakage jet trajectory under each view can be obtained by reverse source tracing fitting. Then, combined with the camera parameters and calibration data of each orthogonal view, the three-dimensional spatial coordinates of the leakage point can be calculated by the spatial intersection algorithm. The accuracy of this unit directly determines the accuracy of the leakage point positioning.

[0067] The annotation generation unit is responsible for mapping the three-dimensional spatial coordinates of the leak point calculated by the three-dimensional positioning unit to the preloaded three-dimensional electronic model and completing the three-dimensional annotation of the leak point. This includes visualizing the location of the leak point on the three-dimensional model and simultaneously annotating the location and attribute information of the leak point. In addition, this unit can also be responsible for generating inspection reports and rework instructions containing three-dimensional annotations, and uploading the pressure data, imaging data, leak point coordinates and annotation results of the entire inspection process to the production line manufacturing execution system to achieve quality data traceability.

[0068] The testing system also includes a data management module, which is an integrated software or hardware unit. Its core function is to systematically collect, process, store, retrieve, and manage all data generated during the testing process. Its role is to ensure the integrity, consistency, and traceability of the testing data, providing a solid foundation for subsequent quality analysis, process optimization, and data interaction with the production line manufacturing execution system. This module can be implemented as a standalone server, an embedded system, or a cloud platform-based application, and is usually combined with a database management system to efficiently store and query data.

[0069] The data management module includes a calibration unit, a model storage unit, and a data storage unit. The calibration unit is specifically used to complete the coordinate calibration between the dual-modal imaging components and the detection space. Its purpose is to establish a unified coordinate mapping relationship among multiple systems, ensuring that the geometric correspondence between different sensors and physical spaces is accurate. Through precise calibration, the three-dimensional spatial coordinates of the leak point can be accurately mapped to the three-dimensional electronic model of the panel under test, thereby achieving high-precision stereo annotation. The calibration process usually involves the use of specific calibration algorithms and calibration tools, such as high-precision calibration boards. By acquiring a series of images with known geometric features, and using methods such as feature point matching, multi-view geometry, or hand-eye calibration, the intrinsic and extrinsic parameters of each imaging component, as well as their transformation matrices with the world coordinate system of the detection space, are calculated.

[0070] The model storage unit is used to pre-store the 3D electronic model corresponding to the panel under test. These 3D electronic models provide detailed geometric information and structural features of the panel under test, serving as a benchmark for 3D annotation of leak points and supporting subsequent leakage point visualization and analysis. The model storage unit can be a database, file system, or dedicated model library, where the stored models can be CAD models, point cloud data, or mesh models, typically containing key information such as the panel's geometry, dimensions, and material properties. Before the testing task begins, the system automatically loads the corresponding pre-stored 3D electronic model based on the model of the panel under test. The data storage unit is the physical or physical storage unit in the data management module responsible for the storage and management of testing data throughout the entire process. The logical storage space records all key data throughout the entire testing process, from initial airtightness inspection to three-dimensional marking of leak points. This includes, but is not limited to, pressure data, infrared imaging data, schlieren imaging data, leak point coordinates, marking results, testing timestamps, and operator information. Through unified storage and management of this data, complete traceability of the testing data can be achieved, providing comprehensive data support for quality analysis, fault diagnosis, and process improvement. The data storage unit can be in the form of hard disks, solid-state drives, network storage devices, or cloud storage. Data can be stored in structured or unstructured form, and corresponding data backup and recovery mechanisms can be configured to ensure data security and reliability.

[0071] By introducing a data management module, this application can systematically manage all data during the testing process. The calibration unit ensures precise coordinate mapping between the dual-modal imaging component and the testing space, enabling the three-dimensional spatial coordinates of the leak point to be accurately mapped onto the three-dimensional electronic model of the panel under test, thereby achieving precise stereoscopic annotation. The model storage unit pre-stores the three-dimensional electronic model of the panel under test, providing an accurate geometric benchmark for the visualization and location of the leak point. The data storage unit uniformly stores and manages the entire process of testing data, including pressure data, various imaging data, leak point coordinates, and annotation results. This not only ensures the integrity and traceability of the testing data but also provides a reliable data source for subsequent quality analysis, process optimization, and data interaction with the production line manufacturing execution system. This comprehensive data management capability significantly improves the accuracy, reliability, and traceability of the testing results, providing a solid technical guarantee for the quality control of the electrical connection panel of the energy storage battery cover.

[0072] After an energy storage battery cover electrical connection panel is assembled on the production line, it is automatically transported to the airtightness testing station. The airtight clamping assembly firmly clamps and seals the panel to be tested, forming a sealed testing chamber. The carbon dioxide supply assembly fills the sealed testing chamber with carbon dioxide gas, and the pressure detection assembly monitors the pressure changes in the sealed testing chamber in real time. If the pressure drops beyond a threshold within a preset time, the panel is determined to be a defective part and requires subsequent precise positioning testing; otherwise, it is determined to be a qualified part and released. This initial inspection method can quickly screen out panels with leaks. Compared with the traditional differential pressure method, which can only determine whether there is an overall leak, this method directly uses carbon dioxide as the filling gas, laying the foundation for subsequent precise positioning testing and avoiding the extra step of gas replacement.

[0073] Before performing precise positioning and detection, the system preloaded a 3D electronic model corresponding to the panel under test. To ensure the accuracy of subsequent imaging and positioning, the system first performed coordinate calibration. By using an infrared and visible light unified calibration board, the system completed the hand-eye joint calibration of the infrared thermal imaging component, three sets of orthogonally arranged schlieren imaging components, and the structured light 3D scanner. This process established a unified transformation matrix between the pixel coordinate system, the camera coordinate system, and the world coordinate system of the detection space. Based on this unified transformation matrix, the system completed the precise coordinate mapping relationship between the 3D electronic model and the detection space, ensuring that all subsequent imaging data and positioning results could be accurately mapped to the 3D structure of the panel.

[0074] For panels deemed unqualified in the initial inspection step, the carbon dioxide supply assembly re-injects carbon dioxide gas into the sealed detection chamber as a tracer gas. The gas pressure introduced this time is higher than that in the initial inspection step to enhance the intensity of the leak jet. The infrared thermal imaging assembly then images the entire area of ​​the panel under test. Because the operating band of the infrared thermal imaging assembly precisely matches the characteristic absorption band of carbon dioxide, it can effectively capture the local temperature or concentration changes caused by carbon dioxide gas leaking from the panel. By analyzing the infrared thermal imaging data, the system can extract the approximate location information of multiple suspected leak areas on the panel. Unlike traditional thermal imaging, which is easily affected by surface reflection and interference from metal materials, this method utilizes the characteristic absorption band of carbon dioxide, significantly improving the accuracy and reliability of detection and reducing false alarms.

[0075] After identifying suspected leak areas, the system no longer performs schlieren imaging on the entire panel. Instead, it targets these suspected leak areas for detection, linking the clamping action of the windproof shield and airtight clamping components. During schlieren image acquisition, the detection space of the panel under test is automatically sealed off, effectively isolating the interference of ambient airflow on the leaking jet and ensuring the purity of the schlieren image. The three orthogonally arranged schlieren imaging components are precisely controlled so that their acquisition field of view is perfectly matched with the coordinate range of the suspected leak area. These three schlieren imaging components simultaneously image the suspected leak area, acquiring schlieren images of the leaking jet from three orthogonal perspectives. This targeted, multi-view, and high-precision imaging method overcomes the limitations of traditional methods in accurately locating leak points, providing high-quality raw data for subsequent precise location.

[0076] After acquiring schlieren images from various orthogonal viewpoints, the image processing unit processes these images. First, a background modeling algorithm is used to filter out static background noise and dynamic environmental interference, extracting a clean foreground image of the leaking jet. Next, the optical flow vector field of the leaking jet in the foreground image is calculated using the optical flow method. Based on the flow direction of the optical flow vector field, the source coordinates of the leaking jet at each viewpoint are obtained through reverse source fitting. If multiple leak points exist, the system performs cluster analysis on multiple sets of jet trajectories extracted from the schlieren images from each orthogonal viewpoint to distinguish the independent jet trajectories of adjacent leak points, and performs reverse source fitting on each set of independent jet trajectories. Finally, the 3D positioning unit accurately calculates the 3D spatial coordinates of the leak point through the source coordinates of the jet at each orthogonal viewpoint and then through spatial intersection fitting. This series of meticulous image processing and 3D positioning steps makes the positioning accuracy of the leak point far exceed that of existing differential pressure methods and traditional thermal imaging methods, achieving precise positioning of ultra-micro leaks.

[0077] The processing module accurately maps the three-dimensional spatial coordinates of the leak point obtained above to the pre-loaded three-dimensional electronic model, completing the three-dimensional annotation of the leak point. In the three-dimensional electronic model, not only is the location information of the leak point annotated, but also information such as leakage intensity and leakage type is simultaneously annotated. The system simultaneously generates a test report with three-dimensional annotation and a detailed rework operation manual to guide production personnel to perform precise repairs. At the same time, the pressure data, imaging data, leak point coordinates and annotation results of the entire testing process are simultaneously uploaded to the production line manufacturing execution system, realizing the traceability of the quality data of the panel under test throughout its entire life cycle.

[0078] Furthermore, for panels under test that have already been tested within the same production batch, the system accumulates and records the leakage point markings based on their three-dimensional markings, and counts the leakage frequency in each preset structural area of ​​the panel under test. When the leakage frequency in any structural area reaches the preset frequency threshold, the system intelligently adjusts the subsequent testing strategy: for the non-conforming parts of the batch that are subsequently determined by the airtightness screening, infrared thermal imaging and / or schlieren imaging are prioritized for the structural areas that reach the preset frequency threshold. After these high-risk areas are inspected, infrared thermal imaging inspection of the entire panel under test is then performed. This intelligent testing strategy significantly improves testing efficiency and can quickly detect batch quality problems.

[0079] Furthermore, based on the coordinates of leakage points, attribute information, and detection data of a batch of panels under test, the system uses cluster analysis to obtain structural regions where leakage is concentrated. Combined with the production process parameters corresponding to these concentrated leakage regions, the system can output process optimization parameters for the panel production process, thus forming a closed-loop process of detection, analysis, and optimization. This method can not only identify problems but also help companies solve problems at the source, continuously improving product quality and production efficiency, which is something that existing technologies cannot achieve.

Claims

1. A method for airtightness testing of the electrical connection panel of an energy storage battery cover, comprising the following steps: S1: The panel to be tested is clamped and sealed to form a closed testing cavity. Gas is injected into the closed testing cavity, and the overall airtightness is initially inspected by detecting the pressure change in the closed testing cavity to distinguish between qualified and unqualified parts. The method is characterized by further including the following steps: S2: Preload the three-dimensional electronic model corresponding to the panel under test, complete the coordinate calibration of at least two sets of orthogonally arranged schlieren imaging systems and the detection space, and establish a unified coordinate mapping relationship between the three-dimensional electronic model, the schlieren imaging system and the detection space. S3: Carbon dioxide gas is introduced into the sealed detection chamber as a tracer gas. For the unqualified parts determined by the initial air tightness inspection, infrared thermal imaging matching the characteristic absorption band of carbon dioxide is used to image the entire area of ​​the panel under test and extract the suspected leakage area. S4: For the suspected leakage area, control the orthogonally deployed schlieren imaging system to simultaneously acquire schlieren images of the leaking jet from each orthogonal viewpoint for the suspected leakage area; S5: Extract the leakage jet trajectory from the schlieren images of each orthogonal viewpoint, obtain the jet source coordinates of the leakage jet trajectory under each viewpoint through reverse source tracing fitting, and obtain the three-dimensional spatial coordinates of the leakage point through spatial intersection fitting of each orthogonal viewpoint. S6: Map the three-dimensional spatial coordinates of the leak point to the three-dimensional electronic model described in step S2 to complete the three-dimensional annotation of the leak point.

2. The airtightness detection method for the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, In step S1, carbon dioxide gas is introduced into the sealed testing chamber for the initial airtightness inspection of the whole system; in step S3, for the unqualified parts determined by the initial airtightness inspection, the gas pressure introduced into the sealed testing chamber is higher than the gas pressure introduced in S1.

3. The airtightness detection method for the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, In step S2, an infrared-visible light unified calibration board is used to complete the hand-eye joint calibration of the infrared imager, the orthogonally arranged schlieren imaging system, and the structured light 3D scanner. A unified transformation matrix of the pixel coordinate system, camera coordinate system, and detection space world coordinate system is established. Based on the unified transformation matrix, the coordinate mapping relationship between the 3D electronic model and the detection space is established.

4. The method for detecting the airtightness of the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, In step S4, the orthogonally arranged schlieren imaging system is configured with three sets of schlieren imaging units. The three sets of schlieren imaging units are arranged with the detection space world coordinate system as the reference. The optical axes of the three sets of schlieren imaging units are orthogonal to each other and are all coplanar with the detection plane of the panel under test. Each set of schlieren imaging units is equipped with independent schlieren optical components and image acquisition units.

5. The method for detecting the airtightness of the electrical connection panel of the energy storage battery cover according to claim 4, characterized in that, In step S4, based on the coordinate range of the suspected leakage area, three sets of schlieren imaging units are controlled to perform targeted imaging acquisition of the suspected leakage area, while schlieren image acquisition is not performed in non-suspected areas. The acquisition field of view of the three sets of schlieren imaging units is simultaneously controlled to be completely matched with the coordinate range of the suspected leakage area.

6. The method for detecting the airtightness of the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, In step S5, a clean foreground image of the leaking jet is extracted from the schlieren image of each orthogonal viewpoint. Then, the optical flow vector field of the leaking jet in the foreground image is calculated by optical flow method. Based on the flow direction of the optical flow vector field, the coordinates of the jet source under each viewpoint are obtained by reverse tracing and fitting.

7. The method for detecting the airtightness of the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, In step S5, when there are multiple leakage points, cluster analysis is performed on multiple sets of jet trajectories extracted from the orthogonal viewpoint schlieren images to distinguish the independent jet trajectories of adjacent leakage points. For each set of independent jet trajectories, reverse source tracing fitting and spatial intersection calculation are performed to obtain the independent three-dimensional spatial coordinates corresponding to each leakage point.

8. The method for detecting the airtightness of the electrical connection panel of the energy storage battery cover according to claim 1, characterized in that, After step S6 is completed, for the panels under test that have completed testing within the same production batch, the results of the three-dimensional marking of their leakage points are cumulatively recorded, and the frequency of leakage in each preset structural area of ​​the panel under test is counted. When the frequency of leakage in any structural area reaches the preset frequency threshold, for the subsequent unqualified parts of the batch that have passed the initial airtightness screening, infrared thermal imaging and / or schlieren imaging detection is performed on the structural areas that have reached the preset frequency threshold. After the structural areas that have reached the preset frequency threshold have been inspected, infrared thermal imaging detection of the entire area of ​​the panel under test is then performed.

9. A system for detecting the airtightness of an electrical connection panel on the top cover of an energy storage battery, comprising an airtight clamping assembly, a pressure detection assembly, and a control module, wherein the airtight clamping assembly is used to clamp and seal the panel to be tested to form a sealed detection cavity, and the pressure detection assembly is connected to the sealed detection cavity to detect pressure changes in the sealed detection cavity to complete an initial screening of overall airtightness, characterized in that... The detection system also includes a carbon dioxide supply component, an imaging component, and a processing module. The carbon dioxide supply component is connected to the sealed detection chamber and is used to fill the sealed detection chamber with carbon dioxide as a tracer gas. The imaging components include an infrared thermal imaging component and at least two sets of orthogonally distributed schlieren imaging components. The operating band of the infrared thermal imaging component matches the characteristic absorption band of carbon dioxide, and its field of view covers the entire area of ​​the panel under test. It is used to image the entire area of ​​the panel under test and output the suspected leak area. The schlieren imaging component is used to simultaneously acquire schlieren images of the leaking jet from various orthogonal viewpoints of the suspected leak area. The processing module processes the infrared imaging data and the schlieren imaging data to extract the leaking jet trajectory. Based on the multi-view leaking jet trajectory fitting, the three-dimensional spatial coordinates of the leak point are obtained. The three-dimensional spatial coordinates of the leak point are then mapped onto the three-dimensional electronic model of the panel under test to complete the stereoscopic annotation of the leak point.

10. The airtightness detection system for the electrical connection panel of the energy storage battery cover according to claim 9, characterized in that, The computation processing module includes: The synchronization control unit is used to synchronize the gas supply action with the imaging acquisition. The image processing unit is used to perform noise reduction and feature extraction on infrared imaging data and schlieren imaging data, and output the leakage jet trajectory. A three-dimensional positioning unit is used to obtain the three-dimensional spatial coordinates of the leak point based on multi-view leak jet trajectory fitting. The annotation generation unit is used to complete the three-dimensional annotation of the leak points.

Images