Pole piece slitting imaging dynamic focusing control method

Abstract

The invention discloses a pole piece slitting imaging dynamic focusing control method, and belongs to the technical field of optical imaging. The spatial position of an imaging module is dynamically adjusted through a multi-dimensional movement module, the module is moved in the pole piece slitting direction in the teaching stage, and a cross-sectional image is collected to generate a reference gradient histogram and a cross-correlation data array which are associated with the position; in the production stage, an image gradient histogram is calculated in real time, and the out-of-focus state is detected through cross-correlation analysis; when the focus is out of focus, the module is controlled to move according to a preset step length and collect an image sequence, optimal focus distribution is matched based on cross-correlation array interpolation, and a focus compensation value is determined; and adjusting the coordinates of the imaging module according to the compensation value to realize dynamic focusing. The method breaks through the installation space limitation of a traditional imaging system, solves the problem of focusing response lag through gradient histogram cross-correlation detection and an interpolation compensation mechanism, and remarkably improves the anomaly detection precision and real-time performance of the slitting process.

Description

Technical Field

[0001] The present invention relates to the technical field of optical imaging, and in particular to a dynamic focus control method for pole piece segmentation imaging. Background Art

[0002] With the widespread application of lithium batteries in new energy, consumer electronics, electric vehicles and other fields, the requirements for product safety, stability and consistency in their manufacturing processes are becoming increasingly stringent. In the battery cell manufacturing process, electrode slitting is a key step, and its slitting quality has a direct impact on subsequent winding, lamination and the performance of the entire battery cell. If there are defects such as foreign matter, falling materials, burrs, and dust on the slitting surface, it will not only cause internal short circuits and capacity degradation, but in severe cases, it may even cause safety issues such as thermal runaway. Therefore, online detection and control of the quality of the slitting surface has become an important means to ensure battery safety performance.

[0003] Currently, the industry typically inspects electrode separation surfaces through manual visual inspection or spot checks. However, this method is labor-intensive, inefficient, subjective, and has a high rate of missed detections, making it difficult to meet the demands of modern high-speed, high-precision intelligent manufacturing. In recent years, with the rapid development of machine vision technology, automatic inspection technology based on visible light imaging has become an important research direction for defect detection on separation surfaces due to its advantages such as intuitive imaging, low cost, easy maintenance, and strong adaptability. Visible light imaging can accurately capture images of electrode separation surfaces, identify different types of defects such as residual foreign matter, protruding burrs, and powder and shavings through image processing and analysis algorithms, and implement automatic classification and alarms, providing a feasible solution for efficient, intelligent, and stable inspection.

[0004] However, inspection technology based on visible light imaging still faces many challenges in practical applications. On the one hand, the physical structure and defect types of the cut surface are complex and diverse, and foreign objects made of different materials appear differently in visible light images, placing higher demands on the resolution, contrast, and image processing algorithms of the imaging system. On the other hand, because the optical system is limited by physical properties such as lens depth of field and light source uniformity, as well as the slight jitter, offset, and bending of the pole piece during transmission, it is difficult to maintain a consistent imaging focal plane on the cut surface, thus affecting the accuracy and stability of the inspection.

[0005] Currently, there are two main focusing solutions: one is to achieve fast and stable focusing through an independent autofocus module, which is suitable for high-speed production scenarios, but its structure is complex, the cost is high, and the installation space requirements are large, making it difficult to promote in space-constrained equipment; the other solution is software focusing based on image features, relying on image texture features to evaluate image clarity and dynamically adjust the focal length. This solution is low-cost and highly flexible, but the focusing time is long and it is highly dependent on target features, which is not conducive to the flexible production needs of frequent product model changes.

[0006] Furthermore, the transmission speed of the pole piece can reach up to 80 meters per minute, and defects on the cut surface are mostly at the micron level. To meet the requirements of such high-precision inspection, the camera's field of view must be controlled within the millimeter range, which places extremely high demands on image acquisition frequency, data processing speed, and overall machine response speed. From loading to slitting and winding, the pole piece passes through multiple rollers. Affected by the coil tension, friction, and equipment structure, the position of the pole piece cut surface in space fluctuates significantly, further increasing the risk of imaging focus drift, resulting in detection blind spots or blurred images.

[0007] Therefore, how to overcome the installation space limitations and focus response lag of existing imaging systems and improve the abnormality detection capability of the electrode slitting process has become a technical problem that needs to be solved urgently. Summary of the Invention

[0008] The main purpose of the present invention is to provide a dynamic focus control method for pole piece cutting imaging, which aims to overcome the installation space limitations and focus response lag problems of existing imaging systems and improve the abnormality detection capability of the pole piece cutting process.

[0009] In order to achieve the above object, the present invention proposes a dynamic focus control method for pole piece segmentation imaging, comprising the following steps: S1. Installing an imaging module including a camera, a lens, a light source, and an optical guide assembly on a multi-dimensional motion module, and dynamically adjusting the spatial position of the imaging module through the motion module; S2. In the teaching phase, controlling the motion module to move along a first direction, collecting multiple electrode cross-section images, and generating a reference gradient histogram and a cross-correlation data array associated with the position; S3. In the production stage, the current image is collected in real time and its gradient histogram is calculated, and a cross-correlation calculation is performed based on the reference gradient histogram to detect the out-of-focus state; S4. When out-of-focus is detected, controlling the motion module to move along a first direction with a preset step length and capturing multiple images, performing interpolation calculation based on the cross-correlation data array, matching the cross-correlation distribution of the current image sequence, and dynamically determining a focus compensation position; S5. Adjust the coordinates of the imaging module according to the compensation position to complete focus control.

[0010] In one embodiment of the present application, the teaching stage in S2 includes: Taking the first direction as the Y axis, continuously collect 50 images with a step size of 3 microns, and each image corresponds to a Y axis coordinate; Calculate the gradient histogram of each image and calculate the cross-correlation coefficient with the reference gradient histogram to form a two-dimensional array of Y-axis coordinate-cross-correlation coefficient and store it.

[0011] In one embodiment of the present application, the preset step size in S4 is 20 microns, and three images are continuously collected in the Y-axis direction.

[0012] In one embodiment of the present application, the interpolation calculation in S4 includes: Perform linear interpolation on the cross-correlation data array in the teaching phase to generate the expansion function EX(t), which is expressed as:

[0013] in, Represents an extension function; represents the cross-correlation coefficient function during the teaching phase; Indicates the Y-axis coordinate; Represents the coordinates of the i-th sampling point in the teaching phase; Indicates the sampling step size during the teaching phase; Indicates the number of samples in the teaching phase; Represents the coordinates of adjacent sampling points; Based on the cross-correlation distribution between the expansion function and the current acquired image sequence, the optimal matching position t is determined by minimizing the objective function.

[0014] In one embodiment of the present application, the objective function is:

[0015] in, Indicates the value of the extended function at the offset position; represents the cross-correlation value of the k-th image in the production stage; Indicates the Y-axis coordinate of the k-th sampling point in the production stage; Indicates the sampling step size of the production stage; Indicates the number of samples taken during the production phase; Indicates the sampling point number.

[0016] In one embodiment of the present application, the optimal matching position t must satisfy the following constraints:

[0017] in, Indicates the final compensation coordinates; Indicates the clear focus coordinates marked during the teaching phase; Indicates the starting coordinates of the current collection.

[0018] In one embodiment of the present application, the control of the motion module in S4 includes: Start from the acceleration section before the preset starting position and move to the sampling position at a constant speed; Each time the sampling position is reached, a trigger signal is sent to the camera to shoot until all sampling quantities are completed.

[0019] In one embodiment of the present application, the optical guiding component includes a reflector for guiding the reflected light irradiated by the light source to the dividing surface to the lens.

[0020] In one embodiment of the present application, the multi-dimensional motion module is an X / Y axis orthogonal motion module, wherein: The first direction is the Y-axis, which is used to adjust the imaging clarity; The second direction is the X-axis, which is used to position the dividing section to the center of the field of view.

[0021] The above technical solution enables real-time focus control of image acquisition during the electrode slicing process. The system can quickly identify when the image is out of focus and automatically restore it to the optimal imaging position. This method, based on the cross-correlation calculation principle of image gradient histograms and combined with interpolation fitting and error minimization matching algorithms, has the significant advantages of high focus accuracy, fast speed, and strong robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The present invention will be described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flow chart of the first embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the following specific embodiments are only used to explain the present invention and do not constitute a limitation of the present invention.

[0024] like Figure 1 As shown, in order to achieve the above purpose, the present invention proposes a dynamic focus control method for pole piece segmentation imaging, comprising the following steps: S1. Installing an imaging module including a camera, a lens, a light source, and an optical guide assembly on a multi-dimensional motion module, and dynamically adjusting the spatial position of the imaging module through the motion module; S2. In the teaching phase, controlling the motion module to move along a first direction, collecting multiple electrode cross-section images, and generating a reference gradient histogram and a cross-correlation data array associated with the position; S3. In the production stage, the current image is collected in real time and its gradient histogram is calculated, and a cross-correlation calculation is performed based on the reference gradient histogram to detect the out-of-focus state; S4. When out-of-focus is detected, controlling the motion module to move along a first direction with a preset step length and capturing multiple images, performing interpolation calculation based on the cross-correlation data array, matching the cross-correlation distribution of the current image sequence, and dynamically determining a focus compensation position; S5. Adjust the coordinates of the imaging module according to the compensation position to complete focus control.

[0025] Specifically, an imaging module is fixed to a mounting fixture. The imaging module includes a camera, a lens, a laser light source, and an optical guide assembly. The optical guide assembly can be a reflector. The laser light source is used to illuminate the cross-section of the electrode piece, and the reflector is used to guide the reflected light from the cross-section of the electrode piece to the lens entrance, allowing the light to accurately enter the camera for imaging.

[0026] The mounting fixture is mechanically secured to a multi-dimensional motion module. The module includes at least two linear motion components for movement in the X- and Y-axes, dynamically adjusting the imaging module's position in space and ensuring its flexible mobility to meet autofocus control requirements.

[0027] During the teaching phase, the multi-dimensional motion module is controlled to move the imaging module along a first direction (i.e., the Y-axis) to a starting position. The module then moves continuously at a preset sampling step size (e.g., 3 microns). Each time the module reaches a specified position, the camera is triggered to capture an image of the electrode cross-section. Each captured image undergoes image gradient analysis, extracting its gradient histogram. The gradient histogram corresponding to each position is cross-correlated with the gradient histogram of the sharpest image to generate a sequence of cross-correlation coefficients associated with the coordinates of each sampling position. This results in a reference gradient histogram and a cross-correlation data array corresponding to each sampling position, which serves as a reference for subsequent production stages.

[0028] During the production phase, the system captures an image of the current electrode cross-section in real time during each imaging process and calculates a gradient histogram for this image. This gradient histogram is then cross-correlated with the reference gradient histogram generated during the teaching phase. If the resulting cross-correlation coefficient falls below a preset threshold, the system determines that the image is out of focus and requires autofocus.

[0029] When out-of-focus is detected, the multi-dimensional motion module is controlled to move forward in a first direction (the Y-axis) at a preset step size (e.g., 20 microns), sequentially capturing multiple images, for example, three images. The cross-correlation coefficient between each captured image and the reference gradient histogram is calculated to form a new cross-correlation sequence. This cross-correlation sequence is then fitted and compared with the cross-correlation data array recorded during the teaching phase. An interpolation function is used to construct a reference function, and a least-squares error matching algorithm is used to calculate the reference position that best matches the current image sequence, thereby determining the optimal focus compensation position.

[0030] After converting the focus compensation position into the actual coordinates of the imaging module, the multi-dimensional motion module controls the imaging module to move to the compensation position, completing an autofocus operation. This method can determine the image clarity in real time during each image acquisition process and quickly restore the focus state when it is out of focus, realizing dynamic focus control during the electrode slicing imaging process.

[0031] The above technical solution enables real-time focus control of image acquisition during the electrode slicing process. The system can quickly identify when the image is out of focus and automatically restore it to the optimal imaging position. This method, based on the cross-correlation calculation principle of image gradient histograms and combined with interpolation fitting and error minimization matching algorithms, has the significant advantages of high focus accuracy, fast speed, and strong robustness.

[0032] In one embodiment of the present application, the teaching stage in S2 includes: Taking the first direction as the Y axis, continuously collect 50 images with a step size of 3 microns, and each image corresponds to a Y axis coordinate; Calculate the gradient histogram of each image and calculate the cross-correlation coefficient with the reference gradient histogram to form a two-dimensional array of Y-axis coordinate-cross-correlation coefficient and store it.

[0033] Specifically, before implementing this method, the imaging module is first fixedly mounted on a mounting fixture. The imaging module includes a camera, a lens, a laser light source, and an optical guide assembly. The laser light source is used to illuminate the cross-section of the pole piece. The optical guide assembly is a reflector that guides light reflected from the cross-section of the pole piece to the lens entrance, allowing the reflected light to accurately enter the lens and ultimately form an image on the camera.

[0034] The mounting fixture is fixed to the multi-dimensional motion module by mechanical connection. The multi-dimensional motion module includes two linear electric slides that can move independently along the X-axis and Y-axis directions, which are used to achieve precise positioning and dynamic movement of the imaging module in a two-dimensional plane.

[0035] During the teaching phase, the multi-dimensional motion module is controlled to move the imaging module along the Y-axis to perform the teaching sampling process. With the Y-axis as the first direction, the module moves continuously in a fixed step size of 3 microns along the Y-axis, capturing 50 images. After each step, the camera automatically captures an image of the cross-section of the electrode at the current imaging position and stores it in a linked format with its corresponding Y-axis coordinate, thus establishing a one-to-one correspondence between 50 sets of images and Y-axis coordinates.

[0036] Each image is then processed to extract its gradient histogram. Each gradient histogram is then cross-correlated with the gradient histogram of a selected image as a reference image to obtain the corresponding cross-correlation coefficient. A mapping relationship is established between each Y-axis coordinate and its corresponding cross-correlation coefficient, ultimately forming a two-dimensional array representing the functional relationship between the Y-axis coordinates and the cross-correlation coefficient. This two-dimensional array serves as basic reference data for the subsequent focusing stage and is stored in the system's internal storage device.

[0037] During the production phase, the system captures a real-time image of the electrode cross-section and calculates its gradient histogram for each electrode imaging task. This image's gradient histogram is then cross-correlated with the reference gradient histogram stored during the teaching phase to determine the cross-correlation coefficient for the current image. If this cross-correlation coefficient is less than the set defocus threshold, the system determines the image is out of focus and initiates the autofocus process.

[0038] During autofocus, the multi-dimensional motion module drives the imaging module to perform multi-point sampling in the Y-axis direction at a set large step size (e.g., 20 microns), capturing three images. The cross-correlation coefficients between each of these three images and the reference gradient histogram are calculated to form a new cross-correlation sequence. This cross-correlation sequence is interpolated and fitted with the two-dimensional array of Y-axis coordinates and cross-correlation coefficients generated during the teaching phase to construct a continuous reference function curve. Based on the minimum error principle, a sliding window matching algorithm is used to calculate the best-fit position of the current sampled image sequence on the reference function curve, thereby calculating the Y-axis position that most closely matches the current cross-correlation distribution.

[0039] The above-mentioned optimal matching position is used as the focus compensation position, and the multi-dimensional motion module is controlled to accurately move the imaging module to the compensation position, thereby completing the automatic focusing operation and ensuring that the image of the pole piece cross section always remains clear in the field of view.

[0040] This technical solution achieves a high-resolution mapping between image clarity and position coordinates by capturing 50 images at a high-precision step size of 3 microns during the teaching phase and constructing a two-dimensional array of Y-axis coordinates and cross-correlation coefficients. Combined with real-time cross-correlation detection and an interpolation fitting matching algorithm during the production phase, the system can quickly respond to out-of-focus detection, calculate the precise focus compensation position, and automatically adjust the position of the imaging module.

[0041] In one embodiment of the present application, the preset step size in S4 is 20 microns, and three images are continuously collected in the Y-axis direction.

[0042] Specifically, during the autofocus process, the multi-dimensional motion module controls the imaging module to move sequentially along the Y-axis at a preset step size of 20 microns, continuously capturing three images. After each image is acquired, image processing is immediately performed to extract its gradient histogram and perform a cross-correlation operation with the reference gradient histogram, generating a set of cross-correlation coefficient sequences of length 3. The system then uses this cross-correlation coefficient sequence to perform a matching process with the two-dimensional array saved during the teaching phase. Interpolation is used to construct the extended cross-correlation function for the teaching phase. Combined with the minimum error matching algorithm, the system determines the reference position coordinates that best match the current cross-correlation sequence.

[0043] Using the above technical solution, the sampling step size in the Y-axis direction is set to 20 microns during the autofocus stage, and the current image clarity change trend is quickly obtained by continuously acquiring three images. Combined with the high-density cross-correlation two-dimensional array established with a step size of 3 microns during the teaching stage, the focus compensation position can be calculated efficiently and accurately through interpolation fitting and error minimization algorithm, thereby achieving fast and precise autofocus control.

[0044] In one embodiment of the present application, the interpolation calculation in S4 includes: Perform linear interpolation on the cross-correlation data array in the teaching phase to generate the expansion function EX(t), which is expressed as:

[0045] in, Represents an extension function; represents the cross-correlation coefficient function during the teaching phase; Indicates the Y-axis coordinate; Represents the coordinates of the i-th sampling point in the teaching phase; Indicates the sampling step size during the teaching phase; Indicates the number of samples in the teaching phase; Represents the coordinates of adjacent sampling points; Based on the cross-correlation distribution between the expansion function and the current acquired image sequence, the optimal matching position t is determined by minimizing the objective function.

[0046] In one embodiment of the present application, the objective function is:

[0047] in, Indicates the value of the extended function at the offset position; represents the cross-correlation value of the k-th image in the production stage; Indicates the Y-axis coordinate of the k-th sampling point in the production stage; Indicates the sampling step size of the production stage; Indicates the number of samples taken during the production phase; Indicates the sampling point number.

[0048] In one embodiment of the present application, the optimal matching position t must satisfy the following constraints:

[0049] in, Indicates the final compensation coordinates; Indicates the clear focus coordinates marked during the teaching phase; Indicates the starting coordinates of the current collection.

[0050] Specifically, adjust the X-direction motion module so that the pole piece cross section is imaged at the center of the camera's field of view; adjust the Y-direction motion module so that the pole piece cross section is imaged clearly; at this time, mark the point as the focus position (click the Confirm Focus button), automatically calculate the gradient histogram (called F-Hist) and save it to a file. After setting the control parameters and camera parameters, start teaching (click the teach button). The teaching method will capture a series of images (such as 50) with a certain Y-axis step size (such as 3 microns), calculate the gradient histograms of these images respectively, and then calculate the cross-correlation coefficient between these histograms and F-Hist. Therefore, one Y-axis coordinate corresponds to one histogram and one cross-correlation coefficient. This pair of corresponding data (called pose-cc array) is saved as a two-dimensional array binding recipe to the file. When the system is running in production, each time an image is taken, the image is first detected for defocus. That is, the gradient histogram of the current image is calculated and the cross-correlation coefficient is calculated with F-Hist. When the coefficient is less than the preset threshold, it is judged to be out of focus, and autofocus is required. Start collecting images according to the set focus parameters, such as a step size of 20 microns, and collect 3 images at Y-axis coordinates of 1.02, 1.04, and 1.06 respectively. Data matching is performed according to the following formula: Formula ① Formula ② Formula ③ Formula ④ Formula ⑤ Formula ⑥ Formula ⑦ In formula ①, T(t) and t are the cross-correlation coefficient and Y-axis coordinate obtained through teaching, respectively. λ is the sampling step size set through teaching, n is the number of samples, and EX(t) is the expansion function obtained by interpolating T(t) within the domain of definition. f(r) is the cross-correlation sequence corresponding to the image collected after defocusing during production, r is the Y-axis coordinate of the sample, m is the number of samples, and θ is the sampling step size. Minimize formula ⑤ under the inequality constraint of formula ⑥ to obtain t; finally, substitute it into formula ⑦ to calculate the current focus position. ,in This is the Y-axis coordinate of the image clearly marked during teaching.

[0051] In one embodiment of the present application, the control of the motion module in S4 includes: Start from the acceleration section before the preset starting position and move to the sampling position at a constant speed; Each time the sampling position is reached, a trigger signal is sent to the camera to shoot until all sampling quantities are completed.

[0052] Specifically, the starting position of the autofocus sampling process is set. To ensure smooth motion and high positioning accuracy during the sampling process, the motion module reserves a certain acceleration distance from the set starting position before performing focus sampling. The motion module starts from this reserved position and accelerates to achieve the set constant sampling speed.

[0053] Once the motion module enters the constant speed section, it continues to move forward at a constant speed along the Y-axis. At each predetermined sampling position, the system uses the built-in position detection module to determine whether the preset coordinates have been reached. Whenever the motion module successfully reaches a sampling position, the control system immediately sends a trigger signal to the camera, instructing it to perform an image acquisition operation.

[0054] After receiving the trigger signal, the camera immediately captures the image of the electrode section and uploads the image data to the processing unit. The motion module then continues to move at a constant speed to the next sampling position and repeats the same steps to acquire images.

[0055] The above sampling and triggering process will continue until the system completes the preset number of sampling. After the sampling is completed, the system terminates the current motion control process and enters the subsequent data processing stage.

[0056] This technical solution significantly improves the smoothness of the motion module's operation and the accuracy of the sampling position by setting an acceleration stage before sampling and controlling image sampling at a constant speed. Furthermore, controlling camera shooting through a position trigger mechanism improves the synchronization between image acquisition and motion control, reducing image blur caused by speed variations or position errors.

[0057] In one embodiment of the present application, the optical guiding component includes a reflector for guiding the reflected light irradiated by the light source to the dividing surface to the lens.

[0058] Specifically, in this embodiment, the optical guide assembly includes a reflector. The reflector is installed in the imaging module, located between the light source and the optical path of the lens. The light source is a high-brightness laser light source, which is used to illuminate the cross-section of the battery electrode. The illumination direction is approximately perpendicular to the cross-section of the electrode, or at a certain angle, to enhance the contour and edge features of the image.

[0059] When light from the laser source strikes the cut surface, it produces diffusely reflected light of a certain intensity. To efficiently direct this reflected light into the lens, a reflector is positioned within the optical path to redirect the light. Specifically, the reflector reflects the reflected light from the cut surface, directing it toward the lens's direction of incidence, thereby enabling the lens to receive a clear imaging signal.

[0060] The reflector's mounting structure is equipped with an adjustable angle rotation mechanism, so that during the equipment debugging process, the angle of the reflector can be fine-tuned to achieve precise calibration of the light path direction, so that the reflected light can accurately enter the center of the lens field of view, ensuring that the image captured by the camera is located in the center of the field of view and is clearly focused.

[0061] This technical solution, using a reflector as an optical guide, effectively adjusts the path of light reflected from the laser light source onto the pole piece's split surface and then directed to the lens. By precisely controlling the light path through the reflector, the image is accurately positioned at the center of the lens' field of view, significantly improving the quality and stability of image acquisition.

[0062] In one embodiment of the present application, the multi-dimensional motion module is an X / Y axis orthogonal motion module, wherein: The first direction is the Y-axis, which is used to adjust the imaging clarity; The second direction is the X-axis, which is used to position the dividing section to the center of the field of view.

[0063] Specifically, in this embodiment, the multi-dimensional motion module is a two-axis linear motion module arranged orthogonally in the X-axis and Y-axis directions. The X-axis motion module and the Y-axis motion module are vertically mounted to form a two-dimensional movable platform, which is used to drive the imaging module mounted thereon to achieve high-precision positioning and movement within a plane.

[0064] The Y-axis motion module, acting as the first-axis movement axis, is mounted on the base frame of the entire motion platform. It is responsible for driving the imaging module for minute position adjustments along the vertical (first) direction. The Y-axis motion module primarily controls the relative distance between the imaging module and the cross-section of the electrode being measured. Adjusting this distance optimizes image clarity, i.e., autofocus. The Y-axis motion module supports submicron stepping accuracy and offers constant-speed motion and position feedback capabilities to meet the stringent requirements for position accuracy and response speed required for image sampling and focus control.

[0065] The X-axis motion module, serving as the second axis of motion, is mounted on the Y-axis motion module's platform, perpendicular to the Y-axis. It drives the imaging module in the horizontal (second) direction to precisely move the electrode section to the center of the lens' field of view. Adjusting the X-axis position ensures the section remains centered within the image acquisition area, facilitating subsequent image processing, focus determination, and image consistency control.

[0066] Through coordinated control of the X-axis and Y-axis directions, the multi-dimensional motion module can achieve precise adjustment of the imaging module in a two-dimensional plane, which not only meets the requirements of segmentation section positioning, but also meets the dynamic compensation requirements of autofocus.

[0067] This technical solution employs multi-dimensional motion modules with orthogonal X and Y axes, respectively used to achieve image centering and clarity adjustment, enabling the system to acquire images with high degrees of freedom and precision. The Y axis controls the focus distance, ensuring the image is always in optimal focus; the X axis controls the lateral position, ensuring that the cross-section of the pole piece is always aligned with the center of the field of view, ensuring a stable and consistent image area, effectively improving imaging quality and automated control.

[0068] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made by using the contents of the present invention description and drawings under the inventive concept of the present invention, or direct / indirect application in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for controlling dynamic focus of electrode segmentation imaging, characterized in that: The following steps are involved: S1. Installing an imaging module including a camera, a lens, a light source, and an optical guide assembly on a multi-dimensional motion module, and dynamically adjusting the spatial position of the imaging module through the motion module; S2. In the teaching phase, controlling the motion module to move along a first direction, collecting multiple electrode cross-section images, and generating a reference gradient histogram and a cross-correlation data array associated with the position; S3. In the production stage, the current image is collected in real time and its gradient histogram is calculated, and a cross-correlation calculation is performed based on the reference gradient histogram to detect the out-of-focus state; S4. When out-of-focus is detected, controlling the motion module to move along a first direction with a preset step length and capturing multiple images, performing interpolation calculation based on the cross-correlation data array, matching the cross-correlation distribution of the current image sequence, and dynamically determining a focus compensation position; S5. Adjust the coordinates of the imaging module according to the compensation position to complete focus control.

2. The method for controlling dynamic focus of electrode segmentation imaging according to claim 1, wherein: The teaching phase in S2 includes: Taking the first direction as the Y axis, continuously collect 50 images with a step size of 3 microns, and each image corresponds to a Y axis coordinate; Calculate the gradient histogram of each image and calculate the cross-correlation coefficient with the reference gradient histogram to form a two-dimensional array of Y-axis coordinate-cross-correlation coefficient and store it.

3. The method for controlling dynamic focus of electrode segmentation imaging according to claim 2, wherein: The preset step size in S4 is 20 micrometers, and 3 images are continuously collected in the Y-axis direction.

4. The method for controlling dynamic focus of electrode segmentation imaging according to claim 1, wherein: The interpolation calculation in S4 includes: Perform linear interpolation on the cross-correlation data array in the teaching phase to generate the expansion function EX(t), which is expressed as: in, Represents an extension function; represents the cross-correlation coefficient function during the teaching phase; Indicates the Y-axis coordinate; Represents the coordinates of the i-th sampling point in the teaching phase; Indicates the sampling step size during the teaching phase; Indicates the number of samples in the teaching phase; Represents the coordinates of adjacent sampling points; Based on the cross-correlation distribution between the expansion function and the current acquired image sequence, the optimal matching position t is determined by minimizing the objective function.

5. The method for controlling dynamic focus of electrode segmentation imaging according to claim 4, wherein: The objective function is: in, Indicates the value of the extended function at the offset position; represents the cross-correlation value of the k-th image in the production stage; Indicates the Y-axis coordinate of the k-th sampling point in the production stage; Indicates the sampling step size of the production stage; Indicates the number of samples taken during the production phase; Indicates the sampling point number.

6. The method for controlling dynamic focus of electrode segmentation imaging according to claim 5, wherein: The optimal matching position t must meet the constraints: in, Indicates the final compensation coordinates; Indicates the clear focus coordinates marked during the teaching phase; Indicates the starting coordinates of the current collection.

7. The method for controlling dynamic focus of electrode segmentation imaging according to claim 1, wherein: The control of the motion module in S4 includes: Start from the acceleration section before the preset starting position and move to the sampling position at a constant speed; Each time the sampling position is reached, a trigger signal is sent to the camera to shoot until all sampling quantities are completed.

8. The method for controlling dynamic focus of electrode segmentation imaging according to claim 1, wherein: The optical guide assembly includes a reflector for guiding the reflected light irradiated by the light source to the dividing section to the lens.

9. The method for controlling dynamic focus of electrode segmentation imaging according to claim 1, wherein: The multi-dimensional motion module is an X / Y axial orthogonal motion module, wherein: The first direction is the Y-axis, which is used to adjust the imaging clarity; The second direction is the X-axis, which is used to position the dividing section to the center of the field of view.

Images