Part tray accurate positioning method

Abstract

The invention provides an accurate positioning method for a part tray, and relates to the technical field of part machining, and the method comprises the following steps: providing a part tray provided with an identification point which is arranged at a preset position on the tray, the part tray comprises a station bottom frame, a station machine base, a feeding station frame, a wedge-shaped guide plate frame, a station alignment pin, a guide insertion frame and a tray positioning pin control structure. Acquiring an image of the part tray through an image acquisition device, and performing preprocessing and feature extraction on the image to identify identification points in the image; and according to the identification points in the image, determining initial position and attitude estimation by combining a preset corresponding relation between the identification points and the actual position and attitude of the part tray. Through combination of image recognition and a bacterial foraging optimization algorithm, high-precision positioning of the part tray is achieved, the part tray is accurately adjusted to the preset position through the mechanical arm adjusting mechanism, and the production efficiency and the machining accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of part processing, and particularly to a precise positioning method for a part pallet. Background Art

[0002] Most traditional positioning methods rely on manual operations or simple mechanical devices, such as manual adjustment, limit block positioning, etc. These methods are easily affected by human factors or mechanical wear during the positioning process, resulting in low positioning accuracy.

[0003] For example, when using a limit block for positioning, due to the manufacturing accuracy of the limit block and the deviation of the installation position, as well as the wear of the part pallet during use, it may lead to the deviation of the positioning position, affecting the accuracy of subsequent processing or assembly.

[0004] In addition, more manual intervention is required, such as manually adjusting the position of the part pallet, checking whether the positioning is accurate, etc., which greatly reduces the production efficiency.

[0005] For example, on a busy production line, if each part pallet needs to be manually positioned and checked, it will not only consume a large amount of human resources, but also may lead to positioning errors due to human negligence or fatigue, affecting the normal operation of the production line. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a precise positioning method for a part pallet to ensure that the part pallet can be accurately and quickly positioned and grabbed.

[0007] To solve the above technical problem, the technical solution of the present invention is as follows:

[0008] In a first aspect, a precise positioning method for a part pallet, the method includes:

[0009] Providing a part pallet provided with identification points, the identification points are arranged at predetermined positions on the pallet, and the part pallet includes a station chassis, a station machine base, a loading station frame, a wedge-shaped alignment plate frame, a station alignment pin, a guiding and plugging frame, and a pallet positioning pin control structure;

[0010] Acquiring an image of the part pallet through an image acquisition device, and performing preprocessing and feature extraction on the image to identify the identification points in the image;

[0011] Based on the identification points in the image, in combination with the preset correspondence between the identification points and the actual position and posture of the part pallet, determining an initial position and posture estimate;

[0012] Calculating and optimizing the initial position and posture estimate through a bacterial foraging optimization algorithm to obtain the actual position and posture of the part pallet;

[0013] According to the actual position and attitude of the part tray, the part tray is adjusted through the adjustment mechanism of the robotic arm. During the adjustment process, the alignment plugging frame is slidably plugged with the wedge-shaped alignment plate frame, and the tray positioning pin is movably plugged with the station alignment pin to adjust the part tray to a predetermined position.

[0014] Furthermore, an image of the part tray is obtained by an image acquisition device, and the image is preprocessed and feature-extracted to identify the identification points in the image, including:

[0015] Control the image acquisition device to capture images according to preset parameters, including frame rate, resolution, and exposure time, and preprocess the captured images, including denoising, enhancing contrast, adjusting brightness and color balance, to obtain processed images;

[0016] On the processed image, use image processing algorithms for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the image.

[0017] Furthermore, on the processed image, use image processing algorithms for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the image, including:

[0018] On the processed image, use the edge detection algorithm for edge feature extraction and set the parameters of the edge detection algorithm;

[0019] Traverse each pixel point of the image, calculate the gradient value of each pixel point, and judge whether the corresponding pixel point is an edge point according to a preset threshold, and finally obtain the image after edge detection;

[0020] On the image after edge detection, use the corner detection algorithm for corner feature extraction and set the parameters of the corner detection algorithm, including window size and threshold;

[0021] Calculate the corner response function value of each pixel point in the image, judge whether it is a corner according to a preset threshold, and record the coordinates of the corner; the corner response function value of each pixel point is where R(X,Y) represents the corner response function value of the image at the pixel point coordinates (X,Y); I X represents the gradient component of the image at the pixel point (X,Y) along the X direction; I Y represents the gradient component of the image at the pixel point (X,Y) along the y direction; k represents a constant;

[0022] Prepare a template image according to the image after edge detection and the recorded corner coordinates, and the template image represents the shape and features of the identification points;

[0023] Take the correlation matching algorithm as a template matching method, traverse each position of the image to be matched, and calculate the similarity between the template image and the image to be matched at the corresponding position;

[0024] For the area where the similarity exceeds the preset threshold, it is determined that there is a marking point matching the template image, and the coordinates and similarity values of the corresponding matching points are recorded to determine and extract the marking points in the image.

[0025] Further, the calculation formula for the similarity between the template image and the image to be matched at the corresponding position is:

[0026]

[0027] Among them, S(x′,y′) represents the similarity between the template image and the image to be matched at the position (x′,y′); x,y represent the pixel coordinates in the template image; W(x,y) represents the weight; T(x,y) represents the gray value of the template image at the pixel coordinates (x,y); I′(x′+x,y′+y) represents the gray value of the image to be matched at the pixel coordinates

[0028] (x′+x,y′+y); T 2 (x,y) represents the square of the gray value of the template image at the pixel coordinates (x,y); I′ 2 (x′+x,y′+y) represents the square of the gray value of the image to be matched at the pixel coordinates (x′+x,y′+y).

[0029] Further, according to the marking points in the image, combined with the preset corresponding relationship between the marking points and the actual position and posture of the part tray, determine the initial position and posture estimation, including:

[0030] For each recognized marking point in the image, extract the coordinate value in the image coordinate system to obtain the coordinate list of the marking points in the image coordinate system;

[0031] Preset the coordinate values of each marking point in the physical coordinate system of the part tray and the corresponding relationship between each marking point and the position and posture of the part tray;

[0032] Match the coordinates of the marking points in the image coordinate system with the preset corresponding relationship, and use the matching result to calculate the rotation angle of the part tray relative to the camera, that is, the posture; use the coordinate values of the marking points in the physical coordinate system and the image coordinate system to calculate the translation amount of the part tray relative to the camera, that is, the position;

[0033] Integrate the posture and position to form the initial position and posture estimation of the part tray relative to the camera.

[0034] Further, the initial position and attitude estimation are calculated and optimized by the bacterial foraging optimization algorithm to obtain the actual position and attitude of the part tray, including:

[0035] In the bacterial foraging optimization algorithm, each bacterium represents a solution, that is, an initial position and attitude estimation of the part tray;

[0036] Set the number of bacteria, search range, step size, chemotaxis steps, reproduction steps, and dissipation and migration probability parameters;

[0037] Each bacterium moves randomly near the current position to explore new solutions, and for the new position of each bacterium, the fitness value of the corresponding bacterium is calculated;

[0038] The bacteria divide according to the fitness value to generate new bacterium individuals, and according to the dissipation probability, some bacteria are randomly selected for dissipation processing, that is, removed from the bacterial population; the new bacterial population after reproduction and dissipation processing is migrated to a new search area;

[0039] Repeat the chemotaxis, reproduction, and dissipation and migration processes until the maximum number of iterations is reached to obtain the final solution, that is, the final position and attitude of the part tray.

[0040] Further, the calculation formula for the fitness value corresponding to the bacterium is:

[0041]

[0042] where F represents the fitness value; n represents the number of observation points; i represents the index variable; w i represents the weight of the i-th observation point; r a,i the abscissa of the i-th observation point in the tray coordinate system; p a,i represents the ordinate of the i-th observation point in the tray coordinate system; θ represents the rotation angle of the tray relative to the world coordinate system; t r 、t p represents the translation parameters of the tray in the world coordinate system; r o,i represents the actual observed abscissa of the i-th observation point in the world coordinate system; p o,i represents the actual observed ordinate of the i-th observation point in the world coordinate system.

[0043] Further, according to the actual position and attitude of the part tray, the part tray is adjusted by the adjustment mechanism of the robotic arm. During the adjustment process, the alignment plug-in frame is slidably inserted into the wedge-shaped alignment plate frame, and the tray positioning pin is movably inserted into the station alignment pin to adjust the part tray to a predetermined position, including:

[0044] Utilize the sliding plug-in characteristics of the alignment plug-in frame and the wedge-shaped alignment plate frame to conduct preliminary guidance on the part tray;

[0045] The alignment plug-in frame slides along a preset track, guiding the part tray to approach a predetermined position;

[0046] When the part tray reaches the predetermined position, the tray positioning pin control and the station alignment pin start to be actively plugged in, so that the tray positioning pin control is inserted into the station alignment pin, and a robotic arm is used to adjust the position and attitude of the part tray, and a locking mechanism is used to lock the part tray in the predetermined position.

[0047] In a second aspect, a computing device includes:

[0048] One or more processors;

[0049] A storage device for storing one or more programs, which when executed by the one or more processors cause the one or more processors to implement the method described above.

[0050] In a third aspect, a computer-readable storage medium stores a program that implements the method when executed by a processor.

[0051] The above solution of the present invention has at least the following beneficial effects:

[0052] By setting identification points on the part tray and using an image acquisition device to acquire images for preprocessing and feature extraction, the positions of the identification points can be accurately identified. Combining the correspondence between the preset identification points and the actual position and attitude of the part tray, the position and attitude of the part tray can be initially estimated. Then, through the bacterial foraging optimization algorithm to calculate and optimize the initial estimate, the positioning accuracy is further improved, ensuring that the part tray can be accurately positioned to the predetermined position.

[0053] This method does not rely on specific mechanical devices or manual operations, but uses image processing and optimization algorithms to achieve automatic positioning. Therefore, it has strong adaptability and flexibility, and can adapt to different types, sizes, and weights of part trays, as well as different working environments and conditions. Through the automated and intelligent positioning process, the time of manual intervention and inspection is reduced, and the production efficiency is improved. At the same time, accurate positioning also avoids production line downtime or rework caused by positioning errors, further improving production efficiency and product quality.

[0054] During the adjustment process, the sliding insertion of the alignment plugging frame and the wedge-shaped alignment plate frame, as well as the movable insertion of the tray positioning pin control and the station alignment pin, provide stable guidance and positioning support for the robotic arm. This enables the robotic arm to more accurately and quickly adjust the position and posture of the part tray, reducing the shaking and deviation during the adjustment process. Brief Description of the Drawings

[0055] Figure 1 It is a schematic flowchart of a method for accurately positioning a part tray provided by an embodiment of the present invention. Detailed Description of the Embodiments

[0056] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0057] As Figure 1 shown, an embodiment of the present invention provides a method for accurately positioning a part tray, and the method includes the following steps:

[0058] Step 1: Provide a part tray provided with identification points, the identification points are set at predetermined positions on the tray, and the part tray includes a station chassis, a station base, a loading station frame, a wedge-shaped alignment plate frame, a station alignment pin, an alignment plugging frame, and a tray positioning pin control structure;

[0059] Step 2: Obtain an image of the part tray through an image acquisition device, and perform preprocessing and feature extraction on the image to identify the identification points in the image;

[0060] Step 3: According to the identification points in the image, combined with the preset corresponding relationship between the identification points and the actual position and posture of the part tray, determine the initial position and posture estimation;

[0061] Step 4: Calculate and optimize the initial position and posture estimation through the bacterial foraging optimization algorithm to obtain the actual position and posture of the part tray;

[0062] Step 5: According to the actual position and posture of the part tray, adjust the part tray through the adjustment mechanism of the robotic arm. During the adjustment process, the alignment plugging frame is slidably inserted into the wedge-shaped alignment plate frame, and the tray positioning pin control is movably inserted into the station alignment pin, and the part tray is adjusted to a predetermined position.

[0063] In an embodiment of the present invention, a part tray provided with identification points is provided, and the tray structure is reasonably designed, including a station chassis, a station base, a loading station frame, a wedge-shaped alignment plate frame, a station alignment pin, a guiding and plugging frame, and a tray positioning pin control structure. The setting of the identification points enables the part tray to be identifiable, providing a basis for subsequent image processing and positioning. At the same time, the structural design of the tray also provides stable support and guidance for the subsequent adjustment of the robotic arm.

[0064] An image of the part tray is obtained by an image acquisition device, and the image is preprocessed and feature-extracted to identify the identification points in the image. This step realizes the automatic identification of the part tray, avoiding the cumbersome and error-prone manual identification and improving the accuracy and efficiency of identification.

[0065] Based on the identification points in the image and in combination with the preset correspondence between the identification points and the actual position and posture of the part tray, an initial position and posture estimation is determined. This step realizes the rapid estimation of the initial position and posture of the part tray by utilizing the information of the identification points.

[0066] The initial position and posture estimation is calculated and optimized by a bacterial foraging optimization algorithm to obtain the actual position and posture of the part tray. This step utilizes an advanced optimization algorithm to refine the initial estimation, improving the positioning accuracy and precision, enabling the part tray to be more accurately positioned to the predetermined position.

[0067] According to the actual position and posture of the part tray, the part tray is adjusted by the adjustment mechanism of the robotic arm. During the adjustment process, the guiding and plugging frame slides and plugs with the wedge-shaped alignment plate frame, and the tray positioning pin control plugs with the station alignment pin movably, adjusting the part tray to the predetermined position. This step realizes the automatic adjustment of the part tray. Through the precise control of the robotic arm and in combination with the guiding and positioning functions of the guiding and plugging frame and the tray positioning pin control structure, the part tray can be stably and accurately adjusted to the predetermined position, improving the production efficiency and product quality.

[0068] In a preferred embodiment of the present invention, in step 1 above, a part tray provided with identification points is provided. The identification points are set at predetermined positions on the tray, and the part tray includes a station chassis, a station base, a loading station frame, a wedge-shaped alignment plate frame, a station alignment pin, a guiding and plugging frame, and a tray positioning pin control structure, and may include:

[0069] In an embodiment of the present invention, identification points are set at predetermined positions on the part tray. These identification points are usually feature points with specific shapes, colors or textures, and are used to provide recognizable information during the image acquisition process. The part tray includes a station chassis, a station base, a loading station frame, a wedge alignment plate frame, a station alignment pin, an alignment plug-in frame, and a tray positioning pin control structure. These structures together constitute a stable and reliable tray system.

[0070] Use an image acquisition device (such as a camera) to obtain images of the part tray. These images may include the tray itself and the identification points. Preprocess the acquired images to remove noise, enhance contrast, etc., to improve the image quality. In the preprocessed images, extract feature points with specific shapes, colors or textures. These feature points are the previously set identification points. Through image recognition algorithms, such as template matching, edge detection, etc., accurately identify the identification points in the images and determine their positions. According to the positions of the identification points in the images, combined with the preset correspondence between the identification points and the actual positions and postures of the part tray, calculate the initial position and posture estimation of the part tray.

[0071] Use optimization algorithms such as the bacterial foraging optimization algorithm to further calculate and optimize the initial position and posture estimation to obtain a more accurate actual position and posture of the part tray. According to the optimized actual position and posture of the part tray, adjust the part tray through the adjustment mechanism of the robotic arm. During the adjustment process, the alignment plug-in frame slides and plugs into the wedge alignment plate frame to provide stable guidance for the adjustment of the robotic arm. At the same time, the tray positioning pin control is movably plugged into the station alignment pin to ensure that the part tray can be accurately positioned to the predetermined position.

[0072] In a preferred embodiment of the present invention, step 2, obtaining an image of the part tray through an image acquisition device, and preprocessing and feature extraction of the image to identify the identification points in the image, may include:

[0073] Step 221, control the image acquisition device to capture images according to preset parameters, including frame rate, resolution, and exposure time, and preprocess the captured images, including denoising, enhancing contrast, adjusting brightness and color balance, to obtain processed images;

[0074] Step 222, on the processed images, use image processing algorithms for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the images.

[0075] In an embodiment of the present invention, an instruction is sent to an image acquisition device (such as an industrial camera) through a built-in control module to set capture parameters, including frame rate (the number of images captured per second), resolution (the pixel size of the image), and exposure time (the time when the camera shutter is open). The image acquisition device captures images according to the received parameters and transmits the captured image data to the processing unit of the machine.

[0076] After receiving the image data, the processing unit first performs denoising processing. This involves applying filtering algorithms (such as Gaussian filtering, median filtering) to reduce the noise in the image and make the image clearer. Then, contrast enhancement is performed by adjusting the pixel value range of the image to increase the contrast of the image and make the identification points more prominent. Then, the brightness and color balance of the image are adjusted. Finally, the preprocessed image is saved as a processed image.

[0077] In step 222, the processing unit applies an edge detection algorithm (such as Canny edge detection) to find regions in the processed image where the pixel values change significantly. These regions correspond to the edges in the image. The result of edge detection is a set of edge pixels, which form the contour of the object in the image. Further, a corner detection algorithm (such as Harris corner detection) is applied to find corners in the processed image. Corners are the points where the pixel values change most significantly in the image and are located at the intersections of the edges of the object. The result of corner detection is a set of corner coordinates, which helps to determine the position of the identification points. The preset identification point template is matched with the processed image. The template matching algorithm (such as normalized cross-correlation) determines the position of the identification points by calculating the similarity between the template and each region in the image. After successful matching, the position coordinates of the identification points are recorded.

[0078] Suppose on an automated assembly line in an automobile manufacturing factory, in order to ensure that the part tray can be accurately aligned to the assembly position, engineers adopt an identification point recognition method based on image processing.

[0079] The engineer sent instructions to the industrial camera on the assembly line through the control module built into the computer, setting the capture parameters of the camera: the frame rate was 25 frames per second to ensure continuous and stable images were captured; the resolution was 2048x1536 pixels to provide high-resolution image details; the exposure time was 8 milliseconds to adapt to the lighting conditions on the assembly line and avoid overexposure or underexposure of the images. The industrial camera started capturing images of the part tray according to the received parameters and transmitted the image data to the processing unit of the computer in real time. After receiving the image data, the processing unit first applied the Gaussian filtering algorithm for denoising, removing random noise and impurities in the image and making the image clearer. Then, the processing unit enhanced the contrast of the image by adjusting the pixel value range of the image, making the identification points on the part tray more prominent in the image. The processing unit adjusted the brightness and color balance of the image to ensure that the overall visual effect of the image met the requirements of subsequent processing. Finally, the processing unit saved the preprocessed image as the processed image.

[0080] The processing unit applied the Canny edge detection algorithm to find regions with significant pixel value changes in the processed image. These regions corresponded to the edges in the image, such as the contour of the part tray and the edges of the identification points. The result of edge detection was a set of edge pixels that clearly outlined the contours of the objects in the image. The processing unit further applied the Harris corner detection algorithm to find corners in the processed image. Corners were the points with the most significant pixel value changes in the image and were located at the intersections of the edges of the objects, such as the four corners of the identification points. The result of corner detection was a set of corner coordinates that accurately indicated the possible positions of the identification points.

[0081] The engineer had previously made an identification point template based on the characteristics such as the shape and size of the identification points. The processing unit matched this preset identification point template with the processed image. The template matching algorithm determined the position of the identification point by calculating the similarity between the template and various regions in the image. When the similarity reached the preset threshold, the processing unit considered the matching successful and determined the accurate position of the identification point. After successful matching, the processing unit recorded the position coordinates of the identification point, and these coordinates would be used for subsequent robotic arm control to ensure that the part tray could be accurately aligned to the assembly position.

[0082] Through fine image capture and preprocessing steps, high-quality image data can be obtained, and the image capture parameters and feature extraction algorithms can be adjusted according to different working environments and part tray types to achieve a wider range of adaptability. The automated image capture and processing process reduces the need for manual intervention, improves the automation level and production efficiency of the production line. Accurate identification point recognition provides precise positioning information for subsequent robotic arm adjustment, ensuring that the part tray can be accurately positioned to the predetermined position.

[0083] In another preferred embodiment of the present invention, in step 222, for the processed image, image processing algorithms are used for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the image, which may include:

[0084] Step 2223, on the processed image, an edge detection algorithm is used for edge feature extraction, and the parameters of the edge detection algorithm are set;

[0085] Step 2224, traverse each pixel point of the image, calculate the gradient value of each pixel point, and determine whether the corresponding pixel point is an edge point according to a preset threshold, and finally obtain the image after edge detection;

[0086] Step 2225, on the image after edge detection, a corner detection algorithm is used for corner feature extraction, and the parameters of the corner detection algorithm are set, including the window size and the threshold;

[0087] Step 2226, calculate the corner response function value of each pixel point in the image, determine whether it is a corner according to a preset threshold, and record the coordinates of the corner; the corner response function value of each pixel point is wherein, R(X,Y) represents the corner response function value of the image at the pixel point coordinates (X,Y); I X represents the gradient component of the image in the X direction at the pixel point (X,Y); I Y represents the gradient component of the image in the y direction at the pixel point (X,Y); k represents a constant;

[0088] Step 2227, according to the image after edge detection and the recorded corner coordinates, prepare a template image, and the template image represents the shape and features of the identification point;

[0089] Step 2228, use the correlation matching algorithm as the method of template matching, traverse each position of the image to be matched, and calculate the similarity between the template image and the image to be matched at the corresponding position;

[0090] Step 2229, for the area where the similarity exceeds the preset threshold, it is determined that there is an identification point matching the template image, and record the coordinates and similarity value of the corresponding matching point to determine and extract the identification points in the image.

[0091] In the embodiment of the present invention, the Canny edge detection algorithm is selected because it can effectively detect the edges in the image while suppressing noise. Set the low threshold and high threshold for edge connection to ensure the continuity and accuracy of the edges. For calculating the image gradient, a 3x3 Sobel operator is selected for gradient calculation in the horizontal and vertical directions.

[0092] Step 2224: For each pixel, calculate its gradient values in the X and Y directions, and use the Sobel operator to calculate the horizontal gradient I X and the vertical gradient I Y . According to a preset threshold, determine whether each pixel is an edge point. If the gradient magnitude is greater than the high threshold, it is marked as a strong edge point. If the gradient magnitude is between the low threshold and the high threshold and is connected to a strong edge point, it is marked as a weak edge point.

[0093] Suppress isolated weak edge points and connect strong edge points to form a complete edge.

[0094] Step 2225: Select the Harris corner detection algorithm because it can effectively identify corners in the image, and set the window size for corner response calculation. Set the threshold for corner response to determine whether it is a corner. The constant k is adjusted according to the image characteristics and takes a value between 0.04 and 0.06.

[0095] Step 2226: For each pixel, calculate its corner response function value R(X,Y). If the response value is greater than the preset threshold, it is marked as a corner, and its coordinates are recorded.

[0096] Step 2227: According to the image after edge detection and the recorded corner coordinates, use an algorithm to generate a template image representing the shape and features of the identification points.

[0097] Step 2228: Select a correlation matching algorithm, such as normalized cross-correlation (NCC), move the template image pixel by pixel on the image to be matched, and calculate the similarity at each position. Use the NCC formula to calculate the similarity value between the template image and the image to be matched at the corresponding position.

[0098] Step 2229: If the similarity value exceeds the preset threshold, it is determined that there is an identification point matching the template image. Record the coordinates and similarity value of the matching point to identify and extract the identification points in the image.

[0099] Suppose an image containing multiple part trays is being processed, and each tray has a unique identification point for positioning.

[0100] Using the Canny edge detection algorithm, the tray contour and the edges of the identification points are successfully identified. On the image after edge detection, the machine uses the Harris corner detection algorithm to accurately find the four corners of the identification points. According to the corner coordinates and edge information, a template image representing the shape of the identification points is generated. The template image is moved pixel by pixel on the image to be matched, and the similarity is calculated through the NCC algorithm. All identification points matching the template are successfully found, and the coordinates and similarity values of each matching point are recorded.

[0101] By combining edge detection and corner detection, fiducial points in an image can be identified more accurately, reducing false positives and missed detections. Automating the identification of fiducial points can reduce manual intervention, improve production efficiency and consistency, and lower labor costs. This method is not only applicable to the identification of fiducial points on part trays but can also be extended to other fields that require image recognition and positioning, such as robot navigation and object grasping.

[0102] In another preferred embodiment of the present invention, the calculation formula for the similarity between the template image and the image to be matched at corresponding positions is as follows:

[0103]

[0104] where S(x′, y′) represents the similarity between the template image and the image to be matched at position (x′, y′); x, y represent the pixel coordinates in the template image; W(x, y) represents the weight; T(x, y) represents the gray value of the template image at pixel coordinate (x, y); I′(x′ + x, y′ + y) represents the gray value of the image to be matched at pixel coordinate (x′ + x, y′ + y); T 2 (x, y) represents the square of the gray value of the template image at pixel coordinate (x, y); I′ 2 (x′ + x, y′ + y) represents the square of the gray value of the image to be matched at pixel coordinate (x′ + x, y′ + y).

[0105] In an embodiment of the present invention, the sizes of the template image T(x, y) and the image to be matched I′(x′, y′) are set. The weight function W(x, y) is initialized, and weights are assigned according to the importance or feature intensity of the pixels in the template image. On the image to be matched I′(x′, y′), with the size of the template image as the window, the window position is moved pixel by pixel. For each window position (x′, y′), the following steps are performed:

[0106] Traverse all pixels (x, y) of the template image and calculate the cumulative sum of W(x, y) × T(x, y) × I′(x′ + x, y′ + y).

[0107] Calculate the sum of the squares of the gray values of the template image ∑ x,y W(x, y) × T 2 (x, y), calculate the sum of the squares of the gray values of the image to be matched at the current window position ∑ x,y W(x, y) × I′ 2 (x′ + x, y′ + y), and calculate the square root of the product of the above two sums of squares as the denominator.

[0108] Use a formula to calculate the similarity between the template image and the image to be matched at the current position (x′, y′), record the similarity S(x′, y′) at each window position, and form a similarity map. Traverse the similarity map to find the regions where the similarity value exceeds a preset threshold. For these regions, it is determined that there are identification points matching the template image, and the coordinates and similarity values of the corresponding matching points are recorded.

[0109] By using the weighted normalized cross - correlation (WNCC) formula, the importance of pixels in the template image is considered, making the matching process more accurate. Even in cases such as illumination changes, noise interference, or image rotation, identification points can be effectively identified. The weight function W(x, y) makes the algorithm more sensitive to key features in the template image and relatively insensitive to non - key features, which helps to accurately find identification points in the presence of complex backgrounds or interfering objects. This method is applicable not only to simple binary template images but also to grayscale or color template images, making the algorithm more flexible in application scenarios and able to meet different requirements. By pre - calculating constant terms such as the sum of the squares of the grayscale values of the template image, repeated calculations in the matching process can be reduced, which helps to improve the matching efficiency, especially when dealing with large images or high - resolution images. Automated identification of identification points can reduce manual intervention, improve production efficiency and consistency. This method is widely used in automated fields such as machine vision, robot navigation, and object grasping, which helps to promote the further development and application of related technologies.

[0110] In a preferred embodiment of the present invention, step 3 above, according to the identification points in the image, combined with the preset corresponding relationship between the identification points and the actual position and pose of the part tray, determining the initial position and pose estimate may include:

[0111] Step 331, for each identified identification point in the image, extract the coordinate values in the image coordinate system to obtain a list of the coordinates of the identification points in the image coordinate system;

[0112] Step 332, preset the coordinate values of each identification point in the physical coordinate system of the part tray and the corresponding relationship between each identification point and the position and pose of the part tray;

[0113] Step 333, match the coordinates of the identification points in the image coordinate system with the preset corresponding relationship. Using the matching results, calculate the rotation angle of the part tray relative to the camera, that is, the pose; using the coordinate values of the identification points in the physical coordinate system and the image coordinate system, calculate the translation amount of the part tray relative to the camera, that is, the position;

[0114] Step 334, integrate the pose and position to form an initial position and pose estimate of the part tray relative to the camera.

[0115] In an embodiment of the present invention, an image containing identification points is received and necessary preprocessing is performed, such as grayscale conversion, binarization, denoising, etc., to improve the recognition accuracy of the identification points. An image processing algorithm (such as edge detection) is applied to identify the identification points in the image. For each identified identification point, its coordinate value (in pixel coordinates) in the image coordinate system is extracted, and the coordinate values of all identified identification points are organized into a list, with each element containing the identification point number and the corresponding coordinate value.

[0116] Step 332, define a physical coordinate system for a part tray, select a corner point or the center point of the tray as the origin, and determine the directions of the X, Y, and Z axes. In the physical coordinate system, a fixed coordinate value is preset for each identification point. These coordinate values are obtained through measurement. Establish the correspondence between each identification point and the position and orientation of the part tray, including the coordinate value of the identification point in the physical coordinate system and the transformation relationship of the identification point relative to the tray orientation (such as the rotation angle).

[0117] Step 333, match the coordinate of the identification point in the image coordinate system with the preset correspondence. This is achieved by comparing the coordinate values.

[0118] Suppose there are m pairs of matching identification points, and the coordinate of each point in the image coordinate system is (x i , y i ), and the coordinate in the physical coordinate system is (x i , y i , z i ). To calculate the rotation angle, a rotation matrix R needs to be solved, which rotates the points in the physical coordinate system to align with the points in the image coordinate system. Methods such as singular value decomposition (SVD) or orthogonal matrix solution are used to solve the rotation matrix R. In the two-dimensional case, the rotation matrix is expressed as where θ is the rotation angle. The rotation matrix is solved by minimizing the error function (such as the least squares method), and the rotation angle θ is extracted from it. Using the coordinate values of the identification points in the physical and image coordinate systems, the translation amount of the part tray relative to the camera needs to be calculated.

[0119] Under the perspective projection model, the relationship between the point (x i , y i ) in the image coordinate system and the point (x i , y i , z i ) in the physical coordinate system can be expressed as where K is the internal parameter matrix of the camera, R is the rotation matrix, and t is the translation vector. By solving this system of equations (using the least squares method), the translation vector t = [t x , t y T .

[0120] Finally, the rotation angle θ and the translation vector t are obtained, which describe the initial pose and position estimation of the part tray relative to the camera.

[0121] Step 334: Integrate the calculated rotation angle and translation amount together to form a complete initial position and pose estimation of the part tray relative to the camera. Output the initial estimation result in an appropriate form (such as numerical value, vector, matrix, etc.).

[0122] Suppose there is a part tray with four identification points (A, B, C, D) attached to it. In the physical coordinate system, the coordinates of these four identification points are (0, 0, 0), (100, 0, 0), (100, 100, 0), and (0, 100, 0) respectively. Now, an image containing these four identification points is taken, and their coordinate values in the image coordinate system are recognized. The coordinate list of the identification points in the image coordinate system is extracted, and A(10, 20), B(110, 20), C(110, 120), and D(10, 120) are obtained. Preset the coordinate values of the identification points in the physical coordinate system and their corresponding relationships, match the coordinates in the image coordinate system with the preset coordinates, and confirm that A corresponds to (0, 0, 0), B corresponds to (100, 0, 0), etc. By comparing the coordinate values in the image coordinate system and the physical coordinate system, calculate the rotation angle of the tray relative to the camera (such as rotating 5° around the Z-axis). Using the perspective projection model, calculate the translation amount of the tray relative to the camera (such as translating 10 mm along the X-axis and 20 mm along the Y-axis). Integrate the rotation angle and the translation amount together to form an initial estimation: the tray rotates 5 degrees around the Z-axis relative to the camera, translates 10 mm along the X-axis, and translates 20 mm along the Y-axis.

[0123] By automatically identifying the identification points and calculating the pose and position of the tray, manual intervention is reduced, and the automation level of the production process is improved. The algorithm can accurately identify the identification points and calculate the pose and position of the tray, avoiding human errors and enhancing accuracy and consistency. This method is applicable to part trays of various shapes and sizes, as well as camera systems with different precisions, supporting diverse applications. The initial pose and position estimation provide an important basis for subsequent image processing, object grasping, path planning, etc. The automatic identification and calculation reduce the dependence on manual operations, reduce costs and time consumption, and improve production efficiency.

[0124] In a preferred embodiment of the present invention, in the above step 4, the bacterial foraging optimization algorithm is used to calculate and optimize the initial position and pose estimation to obtain the actual position and pose of the part tray, including:

[0125] Step 441, in the bacterial foraging optimization algorithm, each bacterium represents a solution, that is, an initial position and attitude estimation of the part tray;

[0126] Step 442, set the number of bacteria, search range, step size, chemotaxis steps, reproduction steps, and dissipation and migration probability parameters;

[0127] Step 443, each bacterium makes a random move near the current position to explore new solutions, and for the new position of each bacterium, calculate the fitness value of the corresponding bacterium;

[0128] Step 444, the bacteria divide according to the fitness value to generate new bacterium individuals, and according to the dissipation probability, randomly select some bacteria for dissipation processing, that is, remove them from the bacterial population; migrate the new bacterial population after reproduction and dissipation processing to a new search area;

[0129] Step 445, repeat the chemotaxis, reproduction, and dissipation and migration processes until the maximum number of iterations is reached to obtain the final solution, that is, the final position and attitude of the part tray.

[0130] In the embodiment of the present invention, in the bacterial foraging optimization algorithm, each bacterium represents a solution, that is, an initial position and attitude estimation of the part tray. This solution can be represented as a vector, including rotation angles (such as rotations around the X, Y, and Z axes) and translation amounts (such as translations along the X, Y, and Z axes). Randomly generate a group of bacteria, and the position and attitude estimation values of each bacterium are randomly selected within a preset search range.

[0131] Step 442, determine the total number of bacteria, which affects the search coverage and computational complexity. Define the upper and lower limits of the rotation angle and translation amount to ensure a reasonable search space. Set the step size for the bacteria to move within the search space, which affects the fineness and speed of the search, and the number of times each bacterium explores new solutions near each position. The number of times the bacteria divide to generate new individuals, and define the probabilities of bacterial dissipation (removal from the population) and migration (movement to a new search area).

[0132] Step 443, each bacterium makes a random move near the current position to generate new solutions. The movement direction and distance are determined by the step size and the upper and lower limits of the rotation angle / translation amount. For the new position of each bacterium, calculate its corresponding fitness value. The fitness value is calculated based on the error between the actual position and attitude estimation value of the part tray. The smaller the error, the higher the fitness.

[0133] Step 444: Based on the fitness value, the bacteria divide to generate new individual bacteria. Bacteria with high fitness are more likely to divide and produce more offspring. Randomly select some bacteria for dissipation processing, that is, remove them from the bacterial population, to increase the diversity of the search and avoid local optima. Migrate the new bacterial population after reproduction and dissipation processing to a new search area to explore a wider solution space.

[0134] Step 445: Repeat the chemotaxis, reproduction, dissipation, and migration processes until the maximum number of iterations is reached (e.g., the fitness value converges) to obtain the final solution, that is, the final estimated values of the position and attitude of the part tray.

[0135] Suppose there is a part tray, and its estimated position and attitude values include the rotation angle θ around the Z-axis and the translation amounts t along the X and Y axes. x ,t y . Set the number of bacteria to 100, the search range to θ ∈ [-30°, 30°], t x ,t y ∈ [-50 mm, 50 mm], the step size is 1 degree / 1 mm, the number of chemotaxis steps is 20, the number of reproduction steps is 5, the dissipation probability is 0.1, the migration probability is 0.2, and the maximum number of iterations is 1000. Randomly generate 100 bacteria, and each bacterium contains random values of θ, t x ,t y . Each bacterium moves randomly near its current position, such as increasing or decreasing θ by 1°, t x ,t y increasing or decreasing by 1 mm, and calculate the fitness value of the new position. Based on the fitness value, the bacteria divide to generate new individuals, randomly select 10% of the bacteria for dissipation processing, migrate the remaining bacteria to a new search area, and repeat the above process 1000 times to obtain the final solution.

[0136] The bacterial foraging optimization algorithm has global search ability, can avoid falling into local optima, and improve the probability of finding the optimal solution. The algorithm can adapt to different search spaces and initial conditions, is applicable to the position and attitude estimation problems of various part trays. The bacteria search and reproduce independently, which is suitable for parallel computing and can improve the computing efficiency. The algorithm can automatically optimize the estimated values of the position and attitude of the part tray, reduce manual intervention, and improve the automation level of the production process.

[0137] In a preferred embodiment of the present invention, the calculation formula for the fitness value corresponding to the bacteria is:

[0138]

[0139] Where F represents the fitness value; n represents the number of observation points; i represents the index variable; w represents the weight of the i-th observation point; r a,iThe abscissa of the i-th observation point in the pallet coordinate system; p a,i Represents the ordinate of the i-th observation point in the pallet coordinate system; θ represents the rotation angle of the pallet relative to the world coordinate system; t r 、t p Represents the translation parameters of the pallet in the world coordinate system; r o,i Represents the actual observed abscissa of the i-th observation point in the world coordinate system; p o,i Represents the actual observed ordinate of the i-th observation point in the world coordinate system.

[0140] In an embodiment of the present invention, obtaining data of n observation points includes coordinates (r a,i , p a,i ) in the pallet coordinate system and actual observed coordinates (r o,i , p o,i ) in the world coordinate system. Set a weight w i for each observation point, and the weight can be determined according to the accuracy, importance or other factors of the observation point. Each bacterium represents a solution, that is, the rotation angle θ of the pallet relative to the world coordinate system and the translation parameters (t r , t p ). Randomly generate a bacterial population within a preset search range, and the values of θ, t r and t p of each bacterium are randomly selected within a reasonable range. For each bacterium, calculate the fitness value F. The fitness value F reflects the matching degree between the bacterial solution (i.e., the position and attitude estimation of the pallet) and the actual observed value. The larger the F value, the higher the matching degree. Repeat the chemotaxis, reproduction, dissipation and migration processes until the maximum number of iterations is reached to obtain the final solution, that is, the final position and attitude estimation value of the pallet.

[0141] Through the bacterial foraging optimization algorithm, the position and attitude of the pallet can be estimated with high precision, improving the accuracy and efficiency of the production process. The algorithm has global search ability, can avoid falling into local optima, and ensure finding the final solution. The algorithm can adapt to different numbers and distributions of observation points, as well as different weight settings, and is applicable to various actual production scenarios. The algorithm can automatically optimize the position and attitude estimation value of the pallet, reduce manual intervention, and improve the automation level of the production process.

[0142] In a preferred embodiment of the present invention, in step 5 above, according to the actual position and attitude of the part pallet, the part pallet is adjusted through the adjustment mechanism of the robotic arm. During the adjustment process, the guiding plug-in frame is slidably inserted into the wedge-shaped guiding plate frame, and the pallet positioning pin hole is movably inserted into the station alignment pin to adjust the part pallet to a predetermined position, which may include:

[0143] Step 551: Use the sliding plug-in characteristics of the alignment plug-in frame and the wedge-shaped alignment plate frame to preliminarily guide the part tray.

[0144] Step 552: The alignment plug-in frame slides along the preset track to guide the part tray closer to the predetermined position.

[0145] Step 553: When the part tray reaches the predetermined position, the tray positioning pin control and the station alignment pin start to actively plug in, so that the tray positioning pin control is inserted into the station alignment pin, and the robotic arm adjusts the position and attitude of the part tray, and uses the locking mechanism to lock the part tray in the predetermined position.

[0146] In the embodiment of the present invention, sensors of the robotic arm (such as vision sensors, laser sensors, etc.) identify the position and attitude of the part tray. According to the recognition result, the robotic arm controls the alignment plug-in frame to slide along the preset track towards the part tray. The sliding plug-in characteristics between the alignment plug-in frame and the wedge-shaped alignment plate frame enable the frame to smoothly and accurately guide the part tray. During the sliding process, the alignment plug-in frame conducts preliminary guidance and fine adjustment on the part tray, making it gradually approach the predetermined position.

[0147] Step 552: The alignment plug-in frame continues to slide along the preset track, maintaining close contact with the part tray. The robotic arm dynamically adjusts the sliding speed and direction of the alignment plug-in frame according to the real-time information fed back by the sensors, ensuring that the part tray can smoothly and accurately approach the predetermined position. When the part tray approaches the predetermined position, the sensors of the robotic arm will send a signal to notify the preparation for the next operation.

[0148] Step 553: When the part tray is near the predetermined position, the tray positioning pin control and the station alignment pin start to actively plug in, and the robotic arm controls the tray positioning pin control to gradually insert into the station alignment pin. During the plugging process, the robotic arm makes fine adjustments to the position and attitude of the part tray to ensure that the tray positioning pin control can be smoothly and accurately inserted into the station alignment pin. After the tray positioning pin control is completely inserted into the station alignment pin, the robotic arm activates the locking mechanism to lock the part tray in the predetermined position. The sensors of the robotic arm confirm again whether the position and attitude of the part tray are correct and feed the result back to the control system. If everything is normal, it enters the next processing process; if there is an abnormality, corresponding adjustments or alarm processing are carried out.

[0149] Suppose there is a part tray that needs to be adjusted to a predetermined position for processing by a robotic arm. The robotic arm is equipped with an alignment plug-in frame, a tray positioning pin control, and a locking mechanism.

[0150] The vision sensor of the robotic arm identifies the position and attitude of the part tray. The robotic arm controls the alignment and insertion frame to slide along a preset track towards the part tray direction, for preliminary guiding and fine-tuning of the tray. The alignment and insertion frame continues to slide along the track, maintaining close contact with the part tray. The robotic arm dynamically adjusts the sliding speed and direction according to the real-time information fed back by the sensor. When the tray approaches the predetermined position, the sensor emits a signal. When the tray reaches near the predetermined position, the tray positioning pin control and the station alignment pin start to engage and insert. The robotic arm controls the tray positioning pin control to gradually insert into the station alignment pin, while making fine adjustments to the tray. When the tray positioning pin control is fully inserted, the robotic arm activates the locking mechanism to lock the tray in the predetermined position. After the sensor reconfirms that the position is correct, it enters the next processing step.

[0151] Through the sliding insertion and active insertion characteristics of the alignment and insertion frame and the tray positioning pin control, high-precision positioning of the part tray is achieved, ensuring the accuracy and consistency of the processing process. The entire process requires no manual intervention and is completely automatically completed by the robotic arm, improving the automation level and efficiency of the production process. This adjustment mechanism can adapt to part trays of different shapes and sizes, as well as different processing stations and process requirements, with strong adaptability and flexibility. The locking mechanism can ensure that the part tray does not move or shake during the processing, improving the stability and reliability of the processing. The entire adjustment process is precisely controlled by the robotic arm and the sensor, avoiding potential safety hazards that may be brought by manual operation and improving the safety of the production process.

[0152] An embodiment of the present invention also provides a computing device, including: a processor, a memory storing a computer program, and when the computer program is run by the processor, it executes the method as described above. All implementation manners in the above method embodiments are applicable to this embodiment and can also achieve the same technical effects.

[0153] An embodiment of the present invention also provides a computer-readable storage medium storing instructions, and when the instructions are run on a computer, the computer is made to execute the method as described above. All implementation manners in the above method embodiments are applicable to this embodiment and can also achieve the same technical effects.

[0154] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A precise positioning method for a part tray, characterized in that, The method includes: Providing a part tray with identification points, the identification points being set at predetermined positions on the tray, and the part tray including a station chassis, a station base, a loading station frame, a wedge alignment plate frame, a station alignment pin, a guiding and plugging frame, and a tray positioning pin control structure; Obtaining an image of the part tray through an image acquisition device, and performing preprocessing and feature extraction on the image to identify the identification points in the image; Based on the identification points in the image, combining with the preset correspondence between the identification points and the actual position and attitude of the part tray, determining an initial position and attitude estimation; Calculating and optimizing the initial position and attitude estimation through a bacterial foraging optimization algorithm to obtain the actual position and attitude of the part tray; According to the actual position and attitude of the part tray, adjusting the part tray through the adjustment mechanism of the robotic arm. During the adjustment process, the guiding and plugging frame is slidably inserted into the wedge alignment plate frame, and the tray positioning pin control is movably inserted into the station alignment pin to adjust the part tray to a predetermined position.

2. The precise positioning method of the part tray according to claim 1, characterized in that, Obtaining an image of the part tray through an image acquisition device, and performing preprocessing and feature extraction on the image to identify the identification points in the image, including: Controlling the image acquisition device to capture an image according to preset parameters, including frame rate, resolution, and exposure time, and performing preprocessing on the captured image, including denoising, enhancing contrast, adjusting brightness and color balance, to obtain a processed image; On the processed image, applying image processing algorithms for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the image.

3. The method for precise positioning of the part tray according to claim 2, characterized in that, On the processed image, applying image processing algorithms for feature extraction, including edge detection, corner detection, and template matching, to identify the identification points in the image, including: On the processed image, applying an edge detection algorithm for edge feature extraction and setting the parameters of the edge detection algorithm; Traversing each pixel point of the image, calculating the gradient value of each pixel point, and judging whether the corresponding pixel point is an edge point according to a preset threshold, finally obtaining the image after edge detection; On the image after edge detection, applying a corner detection algorithm for corner feature extraction and setting the parameters of the corner detection algorithm, including window size and threshold; Calculate the corner response function value of each pixel in the image, determine whether it is a corner according to a preset threshold, and record the coordinates of the corner; the corner response function value of each pixel is where R(X,Y) represents the corner response function value of the image at the pixel coordinates (X,Y); I X represents the gradient component of the image along the X direction at the pixel (X,Y); I Y represents the gradient component of the image along the y direction at the pixel (X,Y); k represents a constant; According to the image after edge detection and the recorded corner coordinates, preparing a template image, the template image representing the shape and features of the identification points; Using the correlation matching algorithm as the method of template matching, traversing each position of the image to be matched, and calculating the similarity between the template image and the image to be matched at the corresponding position; For the area where the similarity exceeds the preset threshold, it is determined that there is an identification point matching the template image, and the coordinates and similarity value of the corresponding matching point are recorded to determine and extract the identification points in the image.

4. The method for precise positioning of the part tray according to claim 3, wherein The calculation formula for the similarity between the template image and the image to be matched at the corresponding position is: Among them, S(x′, y′) represents the similarity between the template image and the image to be matched at the position (x′, y′); x, y represent the pixel coordinates in the template image; W(x, y) represents the weight; T(x, y) represents the gray value of the template image at the pixel coordinates (x, y); I′(x′ + x, y′ + y) represents the gray value of the image to be matched at the pixel coordinates (x′ + x, y′ + y); T 2 (x, y) represents the square of the gray value of the template image at the pixel coordinates (x, y); I′ 2 (x′ + x, y′ + y) represents the square of the gray value of the image to be matched at the pixel coordinates (x′ + x, y′ + y).

5. The precise positioning method for the part tray according to claim 4, wherein Based on the identification points in the image, combining with the preset correspondence between the identification points and the actual position and attitude of the part tray, determining an initial position and attitude estimation, including: For each identification point recognized in the image, extracting the coordinate values in the image coordinate system to obtain a coordinate list of the identification points in the image coordinate system; Preset the coordinate values of each identification point in the physical coordinate system of the part tray and the corresponding relationship between each identification point and the position and attitude of the part tray; Match the coordinate values of the identification points in the image coordinate system with the preset corresponding relationship, and use the matching result to calculate the rotation angle of the part tray relative to the camera, that is, the attitude; use the coordinate values of the identification points in the physical coordinate system and the image coordinate system to calculate the translation amount of the part tray relative to the camera, that is, the position; Integrate the attitude and position to form an initial position and attitude estimation of the part tray relative to the camera.

6. The precise positioning method of the part tray according to claim 5, characterized in that, Calculate and optimize the initial position and attitude estimation through the bacterial foraging optimization algorithm to obtain the actual position and attitude of the part tray, including: In the bacterial foraging optimization algorithm, each bacterium represents a solution, that is, an initial position and attitude estimation of the part tray; Set the parameters of the number of bacteria, search range, step size, chemotaxis steps, reproduction steps, and dissipation and migration probability; Each bacterium moves randomly near the current position to explore new solutions, and for the new position of each bacterium, calculate the fitness value of the corresponding bacterium; The bacteria divide according to the fitness value to generate new bacterium individuals, and according to the dissipation probability, randomly select some bacteria for dissipation processing, that is, remove them from the bacterial population; migrate the new bacterial population after reproduction and dissipation processing to a new search area; Repeat the chemotaxis, reproduction, and dissipation and migration processes until the maximum number of iterations is reached to obtain the final solution, that is, the final position and attitude of the part tray.

7. The precise positioning method for the part tray according to claim 6, characterized in that, The calculation formula for the fitness value of the corresponding bacterium is: Among them, F represents the fitness value; n represents the number of observation points; i represents the index variable; w i represents the weight of the i-th observation point; r a,i is the abscissa of the i-th observation point in the pallet coordinate system; p a,i represents the ordinate of the i-th observation point in the pallet coordinate system; θ represents the rotation angle of the pallet relative to the world coordinate system; t r 、t p represents the translation parameters of the pallet in the world coordinate system; r o,i represents the actual observed abscissa of the i-th observation point in the world coordinate system; p o,i represents the actual observed ordinate of the i-th observation point in the world coordinate system.

8. The method for precise positioning of the part tray according to claim 7, characterized in that, According to the actual position and attitude of the part tray, adjust the part tray through the adjustment mechanism of the robotic arm. During the adjustment process, the guiding plug-in frame is slidably inserted into the wedge-shaped guiding plate frame, and the tray positioning pin control is movably inserted into the station alignment pin to adjust the part tray to a predetermined position, including: Utilize the sliding insertion characteristics of the guiding plug-in frame and the wedge-shaped guiding plate frame to conduct preliminary guiding for the part tray; The guiding plug-in frame slides along the preset track to guide the part tray to approach the predetermined position; When the part tray reaches the predetermined position, the tray positioning pin control and the station alignment pin start to be movably inserted, so that the tray positioning pin control is inserted into the station alignment pin, and use the robotic arm to adjust the position and attitude of the part tray, and use the locking mechanism to lock the part tray at the predetermined position.

9. A computing device, characterized in that, Including: One or more processors; A storage device for storing one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, A program is stored in the computer-readable storage medium, and when the program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images