An object picking method based on correlation analysis and robotic manipulation
By using correlation analysis and computer vision methods, the robot was able to accurately pick up and place objects of different sizes and shapes, solving the problems of insufficient accuracy and adaptability in robot picking operations and improving work efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STATE GRID SHANGHAI MUNICIPAL ELECTRIC POWER CO
- Filing Date
- 2024-07-29
- Publication Date
- 2026-06-23
AI Technical Summary
Existing robots suffer from insufficient precision and poor adaptability in picking up objects, especially when sorting small items and dealing with products of various sizes and shapes, requiring a lot of manual adjustments and affecting work efficiency.
Using a method based on correlation analysis and computer vision technology, the center coordinates of an object are calculated and converted into actual distance coordinates through image acquisition, image processing and target detection, and a six-axis robotic arm is controlled to accurately pick up and place the object.
It improves the robot's automatic adaptability and grasping accuracy when handling objects of different sizes and shapes, simplifies the calculation process, reduces the instruction set, and saves training time.
Smart Images
Figure CN118990472B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of robotics, and in particular to the field of robot object-picking methods. Background Technology
[0002] With the continuous advancement of the industrial intelligence era, the application of robots has expanded from basic tasks such as transportation, inspection, and gripping in industries and service sectors to more advanced applications that replace manual labor in delicate tasks such as fine fingertip manipulation and picking up objects. This demonstrates the rapid development of robot technologies such as precision sensing and precision operation.
[0003] In existing technologies, the precision of robot picking in fine operations may be insufficient due to limitations of sensors and operator errors. For example, when sorting small items, the robot's gripping position may deviate, causing errors. Furthermore, the adaptability of sorting robots to products of various sizes and shapes is also problematic. Addressing these issues requires independent programming and debugging for each product, which typically requires a significant amount of manual labor and considerable time. These processes significantly impact work efficiency and make it difficult to improve efficiency. Summary of the Invention
[0004] The purpose of this invention is to provide an object picking method based on correlation analysis and robot operation. This method optimizes the computer vision algorithm for picking and placing objects by the robotic arm based on correlation analysis, which can greatly improve the robot's automatic adaptability and simplify the calculation process to improve accuracy.
[0005] To achieve the above objectives, the technical solution of the present invention is: an object picking method based on correlation analysis and robot operation, characterized by the following steps:
[0006] Step 1: Obtain the ROI image using an image acquisition device;
[0007] Step 2: Perform image processing and target detection on the ROI image. Specifically, import the ROI image into the software, process it into grayscale format to obtain the ROI grayscale image, and then detect and locate the object image in the ROI grayscale image.
[0008] Step 3: Calculate the center coordinates of the object image center in the ROI grayscale image, and convert the center coordinates into actual distance coordinates in space;
[0009] Step 4: Use the actual distance coordinates in the robot control commands, and the robot will perform actions according to the robot control commands.
[0010] The ROI grayscale image in the second step is the ROI grayscale image of the intensity image. When performing object detection, the object image in the ROI grayscale image is located through two-dimensional cross-correlation operation. The ROI grayscale image is matrix A, and the object image is matrix B. The two-dimensional cross-correlation operation is the discrete cross-correlation between operation matrices A and B, which is given by the following method.
[0011] Matrix
[0012] Among them, 0 ≤ i < ar + ac - 1 & 0 ≤ j < br + bc - 1, a represents the row of matrix A, b represents the column of matrix A, i represents the row of matrix C, j represents the column of matrix C, ar and ac are the dimensions of matrix A, br and bc are the dimensions of matrix B. Evaluate the cross-correlation between the ROI grayscale image and the object image, estimate the correlation degree between the ROI grayscale image and the object image, and store the result in matrix C to obtain the exact position of the object image in the ROI grayscale image, and draw a rectangular box around the exact position in the ROI grayscale image to obtain the ROI processed image, so as to perform positioning.
[0013] In the third step, the center coordinates of the object are calculated through the ROI processed image. Specifically, the processed image is traversed in a loop until the pixel with the minimum brightness is encountered. This pixel with the minimum brightness corresponds to the upper left corner of the rectangular box in the ROI processed image. The center coordinates of the object image in the ROI processed image are calculated and positioned through this pixel. The center coordinates are [x + (a / 2), y + (b / 2)], where x and y are the pixel coordinates of the upper left corner pixel of the rectangular box, and a and b are the dimensions of the rectangular box.
[0014] Converting the center coordinates into actual distance coordinates is obtained by calculating the center coordinates with the image constants of the ROI processed image. Specifically, the origin of the robot is exactly at the middle position of the row of the ROI processed image as the origin. Let the actual distance coordinates be [xdis, ydis]. xdis is the distance from the center coordinates to the row direction of the ROI processed image, ydis is the distance from the center coordinates to the bottom edge of the ROI processed image, x' is the horizontal distance from the origin of the robotic arm to the target object, y' is the pixel distance from the center coordinates to the bottom edge of the ROI processed image, r is the number of pixels in the row of the ROI processed image, c is the number of columns of pixels in the ROI processed image, x' = ±[(x + (a / 2) - r / 2)], y' = c * (y + b / 2). xdis is obtained by multiplying x' by the value of each pixel in the row of the image constants of the ROI processed image, and ydis is obtained by multiplying y' by the value of each pixel in the column of the image constants of the ROI processed image.
[0015] In step four, the robot's motion parameters are calculated using actual distance coordinates. These motion parameters include the rotation angle denoted as θ, the parameter distance denoted as t, and the effective distance required for the lower edge of the ROI denoted as Y. DIS ;
[0016] The following settings illustrate the calculation of motion parameters: the ROI image is 80×110 pixels, the rectangle is 20×20 pixels, the row pixel value of the ROI image constant is 2.25 mm, and the column pixel value of the ROI image constant is 2.27 mm. Therefore, c = 110, r = 80, a = 20, b = 20.
[0017] The column length of the ROI processed image is denoted as C', and the row length of the ROI processed image is denoted as R'. C' is obtained by multiplying c by the value of each pixel in the column of the image constant of the ROI processed image, i.e., C' = 2.27 * c = 249.7. R' is obtained by multiplying r by the value of each pixel in the row of the image constant of the ROI processed image, i.e., R' = 2.25 * r = 180.
[0018]
[0019] Y DIS =ydis–t.
[0020] The robot control commands are for a six-axis robotic arm. The axes of the six-axis robotic arm from its base to its gripping end are sequentially labeled J1, J2, J3, J4, J5, and J6. The steps for the six-axis robotic arm to perform actions are as follows:
[0021] S1, the six-axis robotic arm actuator obtains θ and Y. DIS value,
[0022] S2, the six-axis robotic arm aligns at the origin.
[0023] S3, a six-axis robotic arm based on Y DIS The value moves along this negative y-axis direction.
[0024] S4 and J1 are rotated by an angle relative to θ.
[0025] S5, J4, and J5 are rotated by 90° and θ, respectively.
[0026] S6. Perform six-axis robotic arm movement along the negative z-axis at a constant value.
[0027] S7, the gripper of the six-axis robotic arm's end effector closes and grasps the object.
[0028] S8, the six-axis robotic arm moves along the positive z-axis at a constant value.
[0029] S9. Control the end effector of the six-axis robotic arm to translate to the placement station and place the object.
[0030] By adopting the above technical solution, the beneficial effects of this invention are as follows: The method of this invention uses computer vision technology based on artificial intelligence algorithms and correlation analysis to pick up and place objects with the help of a robotic arm. Specifically, it utilizes image processing technology and intelligent algorithms to calculate the two-dimensional correlation between the object and the captured image to locate the object's position. This method also greatly improves the robot's adaptability through better algorithm optimization. When handling items within a certain range, the robot can automatically adjust to pick up objects of different sizes and shapes, such as circular, cylindrical, mixed shapes, and even deformed and irregular shapes. Furthermore, this method allows for multiple picking up of similar objects from random locations, saving time spent training on picking up objects of different shapes, and simplifying the robot's control instruction set without requiring extensive image mapping technology. In summary, the overall algorithm of this method is simple, occupies little memory space, requires no filters, reduces the instruction set, and has a high picking accuracy, thereby achieving the above-mentioned objectives of this invention. Attached Figure Description
[0031] Figure 1 This is a schematic diagram of the device involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0032] Figure 2 This is a schematic diagram of the interface of the device involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0033] Figure 3 This is a schematic diagram of the robotic arm picking operation process involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0034] Figure 4 This is a schematic diagram illustrating the characteristic steps of image processing and target detection involved in an object picking method based on correlation analysis and robot operation according to the present invention.
[0035] Figure 5 This is a schematic diagram of the object center coordinates involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0036] Figure 6 This is a schematic diagram of the distance and position of the right half of the ROI involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0037] Figure 7 This is a schematic diagram of the angular displacement and translational displacement involved in the object picking method based on correlation analysis and robot operation of the present invention.
[0038] Figure 8 This is a schematic diagram illustrating the actual and expected positions of objects involved in an object picking method based on correlation analysis and robot operation according to the present invention.
[0039] Figure 9 This is a schematic diagram of the actual angle and expected angle of an object involved in an object picking method based on correlation analysis and robot operation according to the present invention. Detailed Implementation
[0040] To further explain the technical solution of the present invention, the following detailed description of an object picking method based on correlation analysis and robot operation will be provided through specific embodiments.
[0041] The basic structural layout of the working system equipment in this embodiment is as follows: Figure 1 As shown, the system includes a robot (in the diagram, a 6-DOF robotic arm (also known as a six-axis robotic arm) is used for picking up and placing objects), a control panel and an image acquisition device (the vision system, specifically an RGB camera with 320x240 pixels used to acquire target images) positioned on the side of the robot. The image acquisition device is mounted on the control panel opposite the robotic arm via a bracket. The positions of the robotic arm and the camera are relatively fixed. The region of interest (ROI) within the camera's field of view is also fixed on the control panel. During object picking and placing, the camera captures ROI images to achieve real-time target detection within the ROI. The object's position is located through the two-dimensional correlation between the object and the captured image, as well as image processing techniques. Then, the pixel coordinates are converted into real-time distance coordinates. With the help of software instructions, the object's position is transmitted to the robotic arm, and the robotic arm is controlled by software (MATLAB in this embodiment) to perform the required object picking / placing operations. The interface of the control system for the above-mentioned equipment in this embodiment can be as follows: Figure 2 As shown, image acquisition, image processing, and all operations related to the robotic arm commands are controlled by a PC. To operate the robot using serial communication, the data transmission baud rate is set to 9600 bits / second, the parity bit should be even, followed by 7 data bits, and then two stop bits. The PC communicates with the robotic arm via an RS232 / 485 serial port. The robotic arm accepts a series of predefined strings in serial form using the serial port. The camera is connected to the PC via a Universal Serial Bus port. The microprocessor unit (MPU) inside the motor drive unit controls the motors of each rotary joint of the robotic arm with the help of digital encoders, which provide angular position feedback from the corresponding motors.
[0042] The operation process steps of the object picking method are as follows: Figure 3 The diagram includes,
[0043] Step 1: Obtain the ROI image through an image acquisition device, that is, capture a single frame of the entire ROI where the object is located through a camera to obtain the ROI image.
[0044] Step 2: Perform image processing and object detection on the ROI image. Specifically, import the ROI image into the MATLAB workspace and store it in grayscale format to obtain the ROI grayscale image, and detect the object image of the object in the ROI grayscale image, and locate the object image in the ROI grayscale image.
[0045] In this step, the ROI grayscale image is the ROI grayscale image of the intensity image. The intensity image is one type of image in grayscale images, which can more prominently show object features. The object detection is as Figure 4 shown. Locate the object image in the ROI grayscale image through two-dimensional cross-correlation operation. Here, set the ROI grayscale image as matrix A and the object image as matrix B. The two-dimensional cross-correlation operation is the discrete cross-correlation between operation matrices A and B, which is given by the following method,
[0046] Matrix
[0047] where \(0\leq i < a_r + a_c - 1\) and \(0\leq j < b_r + b_c - 1\), \(a\) represents the rows of matrix A, \(b\) represents the columns of matrix A, \(i\) represents the rows of matrix C, \(j\) represents the columns of matrix C, \(a_r\), \(a_c\) are the dimensions of matrix A, \(b_r\), \(b_c\) are the dimensions of matrix B. The two-dimensional cross-correlation between the ROI grayscale image and the object image is evaluated as the cross-correlation between the two images, estimate the degree of correlation between them, and store the result in matrix C (the maximum correlation between the two images helps to obtain the exact position of the object image (object) in the ROI grayscale image), and draw a rectangle around the corresponding exact position in the ROI grayscale image to obtain the ROI processed image, store the ROI processed image in a variable and transfer it to the workspace to complete the positioning process.
[0048] Step 3: Calculate the center coordinates of the object image center in the ROI grayscale image, and convert the center coordinates to actual distance coordinates of the actual space.
[0049] In this step, calculate the center coordinates of the object through the ROI processed image. Specifically, use a loop to traverse the processed image until a pixel with the minimum brightness is encountered. This pixel is the first pixel with the minimum brightness located at the upper left corner of the rectangle in the ROI processed image. Locate the center coordinates of the center of the object (object image) through this pixel. The center coordinates are \([x+(a / 2), y+(b / 2)]\), where \(x\), \(y\) are the pixel coordinates of the upper left corner pixel of the rectangle, and \(a\), \(b\) are the dimensions of the rectangle, as Figure 5 As shown. Let the actual distance coordinates be [xdis, ydis], where xdis is the distance from the center coordinate to the row direction of the ROI processed image, ydis is the distance from the center coordinate to the bottom edge of the ROI processed image, x' is the horizontal distance from the origin of the robotic arm to the target object, y' is the pixel distance from the center coordinate to the bottom edge of the ROI processed image, r is the number of row pixels of the ROI processed image, and c is the number of column pixels of the ROI processed image.
[0050] Convert the center coordinate into actual distance coordinates. Specifically, the origin of the six-axis robotic arm is taken as the origin exactly at the middle position of the row of the ROI processed image. This origin is the initial position for the six-axis robotic arm to execute translational or rotational actions according to instructions. This origin divides the ROI processed image into two equal halves, that is, the position of the six-axis robotic arm corresponding to the middle of the row of the ROI processed image. Thus, when positioning the target object, there will be the following three situations when moving along the row of the ROI processed image:
[0051] Situation 1: The center coordinate is located in the left half of the ROI processed image, that is, x+(a / 2)<r / 2. In this case, x+(a / 2) will be less than half of the row of the ROI processed image (in pixels). Therefore, the pixel distance of the center coordinate along the row is: x' = r / 2 - (x+(a / 2)), and x' is specifically the horizontal distance from the origin of the robotic arm to the object on the left.
[0052] Situation 2: The center coordinate is located in the right half of the ROI processed image, as Figure 6 shown, that is, x+(a / 2)>r / 2. In this case, x+(a / 2) will be greater than half of the row of the ROI processed image (in pixels). Therefore, the pixel distance of the center coordinate along the row of the ROI processed image is: x' = (x+(a / 2)) - r / 2, and x' is specifically the horizontal distance from the origin of the robotic arm to the object on the left.
[0053] Situation 3: The center coordinate is located in the middle of the ROI processed image, that is, x+(a / 2)=(r / 2). In this case, x is equal to half of the row of the ROI processed image (in pixels). Therefore, x' = 0.
[0054] According to the above three situations, it can be obtained that x' = ±[(x+(a / 2)-r / 2].
[0055] Along the column of the ROI processed image, the robotic arm moves in the negative y direction. Therefore, y' = c*(y + b / 2).
[0056] xdis is obtained by multiplying x' by the pixel value of each row in the image constant of the ROI processed image, and ydis is obtained by multiplying y' by the pixel value of each column in the image constant of the ROI processed image, thus obtaining the corresponding value of the actual distance coordinates [xdis, ydis].
[0057] In addition, the column length of the ROI processed image is denoted as C', and the row length of the ROI processed image is denoted as R'. C' is obtained by multiplying c by the column pixel value in the image constant of the ROI processed image, and R' is obtained by multiplying r by the row pixel value in the image constant of the ROI processed image.
[0058] The following settings serve as an example to illustrate the calculation of motion parameters in step four: the ROI image is 80×110 pixels, the rectangle is 20×20 pixels, the row pixel value in the image constant of the ROI image is 2.25 mm, and the column pixel value in the image constant of the ROI image is 2.27 mm. Therefore, c = 110, r = 80, a = 20, b = 20, C' = 249.7, and R' = 180.
[0059] Step 4: The actual distance coordinates are used in the robotic arm control commands. The robotic arm performs translational or rotational movements according to the robotic arm control commands. Specifically, the motion parameters of the six-axis robotic arm are calculated using the actual distance coordinates. The axes of the six-axis robotic arm from its base to the gripping end are sequentially denoted as J1, J2, J3, J4, J5, and J6. The motion parameters include the rotation angle denoted as θ, the parameter distance denoted as t, and the effective distance required for the lower edge of the ROI denoted as Y. DIS .
[0060] The translational and angular displacement calculations for the six-axis robotic arm's rotation angle θ along the J1 axis are calculated by measuring the angle of the six-axis robotic arm's movement along the x-direction of the ROI image processing row. The rotation angle of the six-axis robotic arm's J5 axis is equal to the rotation angle of the J1 axis. Figure 7 As shown, the rotation angle θ and parameter distance t are displayed. Subtracting the parameter distance t from the pixel distance, and then using the distance from the center coordinates to the bottom edge of the ROI image, yields the effective distance Y required for the lower edge of the ROI. DIS ,,in this way:
[0061] The rotation angle θ, in degrees and rounded to ten decimal places, is expressed as:
[0062]
[0063] The parameter distance t is given by the equation.
[0064] Effective distance Y required for the lower edge of the ROI DIS =ydis–t.
[0065] After calculating θ and Y DIS Then, this information is transmitted to the six-axis robotic arm with the help of instructions. The six-axis robotic arm receives the instructions and uses these values to perform the necessary tasks, while the z-axis motion remains unchanged.
[0066] The robotic arm control command sequence is sent to the arm drive unit, which instructs the robotic arm to perform the required task. When executing the robotic arm control commands, the six-axis robotic arm rotates to form an arc, placing its end effector directly above the center of the object. The steps of the six-axis robotic arm's movement along the y-axis and rotation along the J1 axis are as follows:
[0067] S1, the six-axis robotic arm actuator obtains θ and Y. DIS value,
[0068] S2, the six-axis robotic arm aligns at the origin.
[0069] S3, a six-axis robotic arm based on Y DIS The value moves along this negative y-axis direction.
[0070] S4 and J1 are rotated by an angle relative to θ.
[0071] S5, J4, and J5 are rotated by 90° and θ, respectively.
[0072] S6. Perform six-axis robotic arm movement along the negative z-axis at a constant value.
[0073] S7, the gripper of the six-axis robotic arm's end effector closes and grasps the object.
[0074] S8, the six-axis robotic arm moves along the positive z-axis at a constant value.
[0075] S9. Control the end effector of the six-axis robotic arm to translate to the placement station and place the object.
[0076] The following content presents the conclusions of the grasping test conducted using the above object picking method:
[0077] Table 1 below shows the θ and Y values of the object at five random locations within the ROI. DIS The value of the position and angle is used to move the robotic arm and grasp the object. This method is particularly suitable for grasping objects with a fixed depth (or height) between the robotic arm and the object. Since the height between the robotic arm and the object is constant, object detection and mapping do not need to obtain the depth information of the object from the robotic arm. Therefore, two-dimensional analysis can meet the high-precision object grasping task.
[0078] Table 1
[0079]
[0080] The joints and end effector of the six-axis robotic arm also have small errors in position and angle. The following errors may occur in the device, as shown in Table 2: position error and angle error:
[0081] 1. Rounding error.
[0082] The six-axis robotic arm rounds the value to the nearest whole number, which can introduce minute errors in the position and angle of the joints and end effectors.
[0083] 2. Errors caused by different image constants
[0084] The image constant for the ROI row is 2.25 mm / pixel, while it is calculated to be 2.27 mm / pixel along the column. This difference between the two constants will introduce small errors in calculating position and angle.
[0085] Table 2
[0086]
[0087] The angle and positional accuracy diagrams of the object at five positions are shown below. Figure 8 and Figure 9 As shown in the figure, the actual values of position and angle basically match the corresponding expected values.
[0088] The above embodiments and figures are not intended to limit the product form and style of the present invention. Any appropriate changes or modifications made by those skilled in the art should be considered as not departing from the patent scope of the present invention.
Claims
1. An object picking method based on correlation analysis and robot manipulation, characterized in that, The steps are as follows: Step 1: Obtain the ROI image using an image acquisition device; Step 2: Perform image processing and target detection on the ROI image. Specifically, import the ROI image into the software, process it into grayscale format to obtain the ROI grayscale image, and then detect and locate the object image in the ROI grayscale image. Step 3: Calculate the center coordinates of the object image center in the ROI grayscale image, and convert the center coordinates into actual distance coordinates in space; Step 4: Use the actual distance coordinates in the robot control commands, and the robot will execute actions according to the robot control commands; The second step, the ROI grayscale image, is the ROI grayscale image of the intensity image. During target detection, object images in the ROI grayscale image are located through two-dimensional cross-correlation operations. The ROI grayscale image is a matrix. The object image is a matrix Its two-dimensional cross-correlation operation is an operation matrix. sum matrix The discrete cross-correlation between them is given by the following method. matrix , where 0≤ < &0≤ <br bc 1, Representation matrix OK, Representation matrix The list, Representation matrix OK, Representation matrix The list, , It is a matrix The dimension of , where br and bc are matrices. The dimension of the matrix is used to evaluate the cross-correlation between the grayscale image of the ROI and the object image, estimate the degree of correlation between the grayscale image of the ROI and the object image, and store the results in a matrix. In this process, the exact location of the object image in the ROI grayscale image is obtained, and a rectangular box is drawn around the exact location in the ROI grayscale image to obtain the ROI processed image, thereby performing localization.
2. The object picking method based on correlation analysis and robot operation as described in claim 1, characterized in that, In step three, the center coordinates of the object are calculated using the ROI processed image. Specifically, the processed image is iterated in a loop until the pixel with the lowest brightness is encountered. This pixel with the lowest brightness corresponds to the upper left corner of the rectangle in the ROI processed image. The center coordinates of the object image center in the ROI processed image are calculated and located using this pixel. The center coordinates are [x+(a / 2), y+(b / 2)], where x and y are the pixel coordinates of the upper left corner pixel of the rectangle, and a and b are the dimensions of the rectangle. The conversion of the center coordinates into actual distance coordinates is calculated by combining the center coordinates with the image constants of the ROI processed image. Specifically, the origin of the robot is located exactly in the middle of the row of the ROI processed image. Let the actual distance coordinates be [xdis, ydis], where xdis is the distance from the center coordinates to the row of the ROI processed image, ydis is the distance from the center coordinates to the bottom edge of the ROI processed image, x' is the horizontal distance from the robot's origin to the target object, y' is the pixel distance from the center coordinates to the bottom edge of the ROI processed image, r is the number of pixels in the row of the ROI processed image, and c is the number of pixels in the column of the ROI processed image. x' = ±[(x + (a / 2) - r / 2], y' = c * (y + b / 2). xdis is obtained by multiplying x' by the row-by-row pixel value of the image constants of the ROI processed image, and ydis is obtained by multiplying y' by the column-by-column pixel value of the image constants of the ROI processed image.
3. The object picking method based on correlation analysis and robot operation as described in claim 2, characterized in that, In step four, the robot's motion parameters are calculated using the actual distance coordinates. These motion parameters include the rotation angle, denoted as... The parameter distance is denoted as t, and the effective distance required for the lower edge of the ROI is denoted as Y. DIS ; The following settings illustrate the calculation of motion parameters: the ROI image is 80×110 pixels, the rectangle is 20×20 pixels, the row pixel value of the ROI image constant is 2.25 mm, the column pixel value of the ROI image constant is 2.27 mm, thus c=110, r=80, a=20, b=20; The column length of the ROI processed image is denoted as C', and the row length of the ROI processed image is denoted as R'. C' is obtained by multiplying c by the value of each pixel in the column of the image constant of the ROI processed image, i.e., C' = 2.27 * c = 249.
7. R' is obtained by multiplying r by the value of each pixel in the row of the image constant of the ROI processed image, i.e., R' = 2.25 * r = 180. , Y DIS =ydis–t.
4. The object picking method based on correlation analysis and robot operation as described in claim 3, characterized in that, The robot control commands are for a six-axis robotic arm. The axes of the six-axis robotic arm from its base to its gripping end are sequentially labeled J1, J2, J3, J4, J5, and J6. The steps for the six-axis robotic arm to perform actions are as follows: S1, six-axis robotic arm driver obtained and Y DIS value, S2, the six-axis robotic arm aligns at the origin. S3, a six-axis robotic arm based on Y DIS The value moves along this negative y-axis direction. S4, J1 according to Rotate by an angle, S5, J4, and J5 are rotated 90° respectively. , S6. Perform six-axis robotic arm movement along the negative z-axis at a constant value. S7, the gripper of the six-axis robotic arm's end effector closes and grasps the object. S8, the six-axis robotic arm moves along the positive z-axis at a constant value. S9. Control the end effector of the six-axis robotic arm to translate to the placement station and place the object.