Control method and equipment of tomato picking robot, medium and product

Abstract

The invention discloses a control method and device of a tomato picking robot, a medium and a product, and relates to the technical field of picking robos.The method comprises the steps that a to-be-recognized image is input into a tomato mature fruit recognition model, and a target in the to-be-recognized image is recognized; coordinates in a base coordinate system of the tomato picking robot are determined based on the conversion matrix, path planning is conducted on a chassis of the tomato picking robot through an A * algorithm, and a planned path is obtained; and the walking mechanism is controlled to walk along the planned path, the tail end of the mechanical arm is moved to the target pose, and the tomato picking robot is controlled to clamp, shear and store the target. The control precision of the tomato picking robot is improved through the tomato mature fruit recognition model and the path planning algorithm.

Description

Technical Field

[0001] The present application relates to the technical field of picking robots, and in particular to a control method, equipment, medium and product of a tomato picking robot. Background Art

[0002] With the continuous growth of the global population, agricultural production is facing greater and greater challenges, especially in the context of labor shortage and rising labor costs. Traditional agricultural production methods cannot meet the needs of efficient, precise and large-scale agriculture. To this end, agricultural automation and intelligence have become an important direction for the development of modern agriculture. The application of robotics, artificial intelligence, sensor technology and automatic control technology has provided new solutions for the automated harvesting of crops. At present, picking robots for fruits and vegetables such as tomatoes are gradually becoming an important research direction in the field of agricultural automation. Picking tomatoes is difficult, mainly due to the complexity of the fruit position, the weak nature of the fruit, different planting methods and different lighting conditions. Traditional tomato picking methods mostly rely on manual operation, which has problems such as high labor intensity, low efficiency and easy damage to the fruit. Therefore, the development of an efficient, precise and safe tomato picking robot has become the key to solving these problems.

[0003] The existing tomato picking robot technology has made some progress, but there are still many technical difficulties, which are mainly reflected in the following aspects: 1. Insufficient image recognition accuracy: Some existing picking robots use traditional computer vision technology to detect and locate tomatoes through feature extraction and template matching. At the same time, most of the existing picking robots are in the laboratory research and development stage, and the environment is usually simple and the light is sufficient. However, in complex farmland environments, these methods often have low recognition accuracy due to changes in lighting, cluttered backgrounds, or changes in fruit morphology. Although deep learning technologies (such as YOLO, Faster R-CNN, etc.) have made significant progress in object detection, in agricultural picking scenarios, due to the limitations of data sets and the diversity of crops, the generalization ability and robustness of the model still need to be further improved. 2. Insufficient end control accuracy: The end control of the robotic arm plays a vital role in the picking process. Traditional robotic arm control methods mostly rely on simplified geometric models or image-based positioning, which results in the end of the robotic arm being unable to accurately reach the target position in a changing environment. Especially when soft fruits such as tomatoes need to be picked accurately, insufficient control accuracy may cause fruit damage or picking failure. 3. Communication stability and scalability issues: Since there are often many unknown interference factors in the agricultural environment, many existing robot systems have major communication problems. Especially in scenarios using wireless communication technology, signal instability and transmission delay issues will affect the robot's real-time response capabilities and collaborative control efficiency. In addition, as the functions of the robot system expand and the number of devices increases, how to ensure the system's communication stability, reliability, and good scalability has become an important technical problem. Summary of the invention

[0004] The purpose of this application is to provide a control method, device, medium and product of a tomato picking robot, which can improve the control accuracy of the tomato picking robot.

[0005] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a control method for a tomato picking robot, wherein the control method for a tomato picking robot is applied to a tomato picking robot, wherein the tomato picking robot comprises: A main controller, and an environmental information acquisition module, a walking mechanism, a storage bin and a mechanical arm connected to the main controller; The environmental information acquisition module is used to obtain environmental data required for navigation and obstacle avoidance; The end of the mechanical arm is provided with an image acquisition module, a gripper and a pneumatic scissors; The image acquisition module, the clamp and the pneumatic scissors are all connected to the main controller; The control method of the tomato picking robot comprises: Acquire the environment image in front of the robot arm as the image to be recognized; The image to be identified is input into a tomato ripe fruit recognition model to identify the target in the image to be identified and the coordinates of the target in the image acquisition module coordinate system; the target is a tomato ripe fruit; the tomato ripe fruit recognition model is obtained by training a YOLO model using multiple target images; Based on the transformation matrix and the coordinates of the target in the image acquisition module coordinate system, the coordinates in the tomato picking robot base coordinate system are determined; Taking the coordinates of the tomato picking robot base coordinate system as the end point, the A* algorithm is used to plan the path of the tomato picking robot chassis to obtain the planned path; Controlling the walking mechanism to walk along the planned path, and obtaining the target posture when reaching the end point of the planned path; Moving the end of the robotic arm to a target position; Controlling the clamp and the pneumatic scissors to perform clamping and shearing operations on the target; The end of the robotic arm is controlled to clamp the target to the storage bin for storage.

[0006] Optionally, the image acquisition module is a binocular camera.

[0007] Optionally, the environmental information acquisition module includes: a depth camera and a laser radar.

[0008] Optionally, before acquiring the image of the environment in front of the robot arm as the image to be recognized, the method further includes: Get multiple images containing targets; Whether the target is present and the coordinates of the target in the coordinate system of the tomato picking robot base are used as labels to annotate multiple images containing the target; With the target image as input and the label of the target image as output, the YOLO model can be trained using the loss function to obtain a tomato ripe fruit recognition model.

[0009] Optionally, the loss function is: ; in, is the total loss; is the position error weight coefficient; is the number of grids; is the number of bounding boxes of the i-th grid; is the indicator variable of the jth bounding box of the i-th grid; is the actual frame center coordinate; is the center coordinate of the prediction box; is the actual frame width; is the prediction box width; is the actual frame height; is the predicted box height; are other error terms.

[0010] Optionally, the conversion matrix includes: a conversion matrix from the image acquisition module coordinate system to the robot end coordinate system, and a conversion matrix from the robot end coordinate system to the tomato picking robot base coordinate system.

[0011] Optionally, moving the end of the robotic arm to a target posture includes: The end of the robot arm is moved to detect whether there is an obstacle in front of the end of the robot arm until the end of the robot arm moves to the target posture, and when an obstacle is detected, the movement of the robot arm is stopped.

[0012] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the control method of the above-mentioned tomato picking robot.

[0013] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the control method of the above-mentioned tomato picking robot.

[0014] In a fourth aspect, the present application provides a computer program product, including a computer program, which implements the control method of the above-mentioned tomato picking robot when executed by a processor.

[0015] According to the specific embodiments provided in this application, this application discloses the following technical effects: The present application provides a control method, device, medium and product for a tomato picking robot, which completes the control of the tomato picking robot based on deep learning, A* algorithm path planning, Lie group Lie algebra posture solution and efficient and stable communication. The YOLO deep learning model is introduced for efficient target detection and positioning, and the A* algorithm is combined to achieve efficient path planning. The Lie group Lie algebra is used to control the end of the robot arm. An efficient and stable communication solution based on CAN bus and network cable communication is designed, which can significantly improve the picking accuracy, path planning efficiency, communication stability and system reliability, solve the problems of low efficiency, poor accuracy and poor safety in the prior art, and has broad application prospects and practical value. By integrating modern artificial intelligence, robot control, sensor technology and communication technology, a more efficient, accurate and stable automatic tomato picking solution is provided to meet the urgent needs of modern agricultural production for automation and intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is a flow chart of a control method for a tomato picking robot in one embodiment of the present application.

[0018] Figure 2 This is a schematic diagram of a control method for a tomato picking robot in one embodiment of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0020] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0021] In an exemplary embodiment, a control method of a tomato picking robot is provided, and the control method of the tomato picking robot is applied to a tomato picking robot, and the tomato picking robot includes: a main controller, and an environmental information acquisition module, a walking mechanism, a storage bin and a mechanical arm connected to the main controller. The environmental information acquisition module is used to obtain environmental data required for navigation and obstacle avoidance. An image acquisition module, a clamp and a pneumatic scissors are provided at the end of the mechanical arm. The image acquisition module, the clamp and the pneumatic scissors are all connected to the main controller. Among them, the environmental information acquisition module includes: a depth camera and a laser radar. The image acquisition module is a binocular camera. In this application, wired communication is adopted between the main control and the chassis, relay, mechanical arm and laser radar to ensure the stability and reliability of the system. However, if a larger range of operations and collaboration of multiple robots are considered, a wireless communication solution can be considered. Using Wi-Fi or 5G wireless network technology, multiple robots can share task information and status through wireless communication, realize collaborative operation and improve work efficiency. Although wireless communication may encounter bandwidth and signal interference problems, it is still an effective alternative in specific operating environments. The present invention adopts a crawler chassis, which adapts to uneven terrain and provides higher stability. However, in some flat, hardened ground environments, a wheeled chassis can also be used as an alternative. The wheeled chassis has a higher moving speed and better maneuverability, and is suitable for simpler environments. Although the crawler chassis has more advantages in complex terrain, the wheeled chassis can also complete the same picking task in some application scenarios.

[0022] like Figure 1 As shown, the control method of the tomato picking robot includes: Step 101: Acquire the environment image in front of the robot arm as the image to be recognized.

[0023] The tomato picking robot collects environmental information through the Livox Mid360 LiDAR and ZED 2i depth camera installed on the chassis to obtain the environmental data required for navigation and obstacle avoidance. At the same time, the binocular camera installed at the end of the (6-DOF series) robotic arm obtains close-up tomato images, providing high-resolution image data for subsequent target recognition. All data is transmitted to the main controller (NVIDIA TX2) for analysis and processing.

[0024] Step 102: Input the image to be identified into the tomato ripe fruit recognition model to identify the target in the image to be identified and the coordinates of the target in the image acquisition module coordinate system. The target is a tomato ripe fruit. The tomato ripe fruit recognition model is obtained by training the YOLO model using multiple target images.

[0025] The main controller (TX2) calls the image processing module based on the YOLO (You Only Look Once) model for target recognition. The YOLO model uses a convolutional neural network (CNN) to achieve single-stage detection, which can quickly and accurately identify tomatoes in the image and mark their position and size. The loss function of the model is used to optimize the detection accuracy of the model, which mainly consists of three parts: position error, confidence error, and category error. Before step 101, it also includes: obtaining multiple images containing targets. Whether the target is contained and the coordinates of the target in the coordinate system of the base of the tomato picking robot are used as labels to mark multiple images containing targets. With the target-containing image as input and the label of the target-containing image as output, the YOLO model can be trained using the loss function to obtain a tomato ripe fruit recognition model.

[0026] The loss function is: .

[0027] in, is the total loss, which measures the difference between the predicted and actual results; is the position error weight coefficient, which is used to adjust the importance of the position error in the total loss; is the number of grids; is the number of bounding boxes of the i-th grid; is the indicator variable of the jth bounding box of the i-th grid. If the jth bounding box of the i-th grid detects an object, the value is 1, otherwise it is 0; is the actual frame center coordinate; is the center coordinate of the prediction box; is the actual frame width; is the prediction box width; is the actual frame height; is the predicted box height; is other error term. Among them, the position error weight coefficient The value is usually large, ensuring that the position error accounts for a higher proportion of the total loss, thereby improving positioning accuracy. In addition, other terms in the loss function are used to measure the category confidence error and category error, helping the model improve recognition results under multiple categories and target numbers.

[0028] Step 103: Based on the transformation matrix and the coordinates of the target in the image acquisition module coordinate system, determine the coordinates in the tomato picking robot base coordinate system. The transformation matrix from the image acquisition module coordinate system to the manipulator end coordinate system, and the transformation matrix from the manipulator end coordinate system to the tomato picking robot base coordinate system.

[0029] After identifying the position of the tomato, the coordinate transformation method of Lie group and Lie algebra is used to calculate the exact position of the target relative to the base of the robot arm. The transformation matrix from the image acquisition module coordinate system to the robot end coordinate system : Obtained through prior calibration, this matrix represents the position of the camera coordinate system relative to the end of the robotic arm; the position of the target relative to the image acquisition module That is, the target pose: obtained using the camera's visual depth data, representing the pose of the target tomato in the camera coordinate system; the transformation matrix from the robot end coordinate system to the tomato picking robot base coordinate system :According to the current joint angle of the robot arm, the position of the end of the robot arm relative to the base is obtained by inverse solution.

[0030] The position and orientation of the target in the base coordinate system are calculated by the following transformation: .

[0031] in, Represents the pose of the target tomato in the base coordinate system of the tomato picking robot, including position and orientation.

[0032] Step 104: Taking the coordinates of the tomato picking robot base coordinate system as the end point, use the A* algorithm to perform path planning on the chassis of the tomato picking robot to obtain a planned path.

[0033] After obtaining the target position, the chassis is path-planned to ensure that the robot moves safely and quickly to the optimal picking position. The path planning uses the A* algorithm, and the heuristic function formula of the algorithm is as follows: f(n)=g(n)+h(n).

[0034] Among them: f(n) is the comprehensive cost of node n; g(n) is the actual cost from the starting point to the current node; h(n) is the heuristic cost, which estimates the minimum cost from the current node to the target node (usually using Euclidean distance or Manhattan distance).

[0035] The A* algorithm gradually approaches the target by giving priority to nodes with smaller f(n) values ​​in path search, thereby generating the shortest path. This path planning scheme provides a stable and reliable path reference for chassis motion, avoids obstacles and optimizes the robot's motion path. In addition to using the A* algorithm for chassis path planning, the RRT (Rapid Random Tree) algorithm and the DLite (an incremental path planning algorithm) algorithm can also be used as alternatives. The RRT algorithm is suitable for high-dimensional complex environments and can quickly generate paths, especially in dynamic environments, where it has certain advantages. The DLite algorithm can quickly update the path when the environment changes, and is particularly suitable for handling dynamic obstacles in complex environments. Therefore, in some complex or dynamic environments, the RRT or DLite algorithm can replace the A* algorithm to achieve the function of path planning.

[0036] Step 105: Control the walking mechanism to walk along the planned path, and obtain the target posture when reaching the end of the planned path.

[0037] Step 106: Move the end of the robotic arm to the target posture. Move the end of the robotic arm and detect whether there is an obstacle in front of the end of the robotic arm until the end of the robotic arm moves to the target posture. When an obstacle is detected, stop moving the robotic arm.

[0038] After obtaining the base coordinates of the target, the main controller transmits the target posture information to the robot control module, driving the 6-DOF robot to move the end to the target posture. The robot detects its status in real time during the movement, and detects the angle, torque and end collision of each joint to ensure the safety and accuracy of the robot movement during the picking process. If a collision or abnormality is detected, the system will immediately suspend the movement of the robot and alarm.

[0039] Step 107: Control the gripper and the pneumatic scissors to perform gripping and shearing operations on the target.

[0040] When the robot reaches the target position, the main controller activates the gripper control module and the shear control module, and controls the pneumatic gripper and pneumatic scissors at the end through relays, as follows: 1. Clamping operation: The main controller controls the relay to turn on the clamp solenoid valve, activate the pneumatic gripper, and adjust the clamping force to safely grasp the tomatoes and avoid damaging the fruit skin.

[0041] 2. Cutting operation: If the tomato stem needs to be cut off, the main controller controls the relay to drive the solenoid valve, activating the pneumatic scissors to cut the stem, thereby completing the picking action. During the cutting operation, the main controller monitors the status of the pneumatic scissors in real time to ensure accurate and smooth cutting action.

[0042] Step 108: Control the end of the robot arm to clamp the target to a storage bin for storage.

[0043] After picking is completed, the main controller controls the robot arm to move the gripped tomatoes from the target location to the storage area on the chassis. The robot arm control module adjusts the gripper opening and closing to store the tomatoes in the designated location, ready for the next picking task. The specific steps are as follows: Figure 2 shown.

[0044] This application realizes a high-efficiency, high-precision and high-safety automatic tomato picking process through the comprehensive application of advanced target recognition, path planning, posture control, collision detection and efficient and stable communication technology, significantly improves the performance of the picking robot in complex environments, and enhances the stability, reliability and scalability of the system: 1. Efficient target recognition and positioning: The image recognition method based on the YOLO model is adopted. Through the efficient calculation of deep neural networks and the single-stage detection architecture of YOLO, the rapid recognition and accurate positioning of tomatoes can be achieved in a short time. Compared with traditional algorithms based on feature extraction and image segmentation, the YOLO model can detect multiple targets in real time in complex environments, significantly improving the picking efficiency. Through the refined design of the loss function, the model can accurately locate the position and size of tomatoes to ensure the accuracy of the picking action. 2. Efficiency and stability of path planning: The A* (Astar) algorithm is used for path planning in chassis navigation. The A* algorithm uses a heuristic function to evaluate the path cost, enabling the robot to find the shortest path in a complex farmland environment, avoid obstacles and complete path planning at the lowest cost. This planning method effectively reduces the energy consumption of the system, avoids unnecessary path exploration, and greatly shortens the time to reach the target point, thereby improving the work efficiency of the overall system. 3. Accurate terminal posture control: During picking, Lie group Lie algebra is used to calculate and integrate multiple matrix transformations. This matrix operation ensures that the end of the picking robot arm can accurately move to the location of the tomato and maintain the correct posture, thereby achieving damage-free picking. Compared with the manual calibration or single angle adjustment of the traditional method, the tomato picking robot control method provided in this application reduces the positioning error and improves the accuracy of picking. 4. Efficient collision detection and safety control: A collision detection algorithm is added to the robot arm motion control, which can detect abnormal collisions in the operation of the robot arm in real time. Once a collision or other abnormal state is detected, the control system immediately stops the movement of the robot arm to ensure the safety of the picking equipment and plants. This detection method is more intelligent than the traditional directional control mechanism and can maintain high safety in complex and multi-obstacle environments. 5. Stable and efficient communication system: Two different technical means are used in the communication part to ensure efficient and reliable data transmission: CAN communication is used between the main control and the relay and chassis. The main control communicates with the relay and chassis through the CAN (Controller Area Network) bus, which has high communication stability and real-time performance. The CAN bus has the advantages of strong anti-interference ability, reliable communication, and low latency. It is particularly suitable for applications that require high-frequency communication and large-scale distributed systems. In addition, the CAN bus has strong scalability and can flexibly add more devices without affecting system performance, ensuring the stable operation of the robot in complex environments. The CAN bus has an error reset function. When a communication error is detected, the communication system can be reset, which enhances the reliability of the system.The main control, the robot arm and the laser radar use Ethernet to communicate. The main control, the robot arm and the laser radar use Ethernet for high-speed data transmission, which has the characteristics of wide bandwidth and low latency. This communication method ensures the stability of large data transmission, such as real-time robot arm information, laser radar point cloud information, etc., and ensures that the system can obtain and process a large amount of sensor data in real time. Ethernet communication also has strong anti-interference ability and high bandwidth, which enables the system to operate stably at a higher data transmission rate, especially suitable for control systems requiring high bandwidth and high reliability. 6. Stability and adaptability of the overall system: This application design is based on the NVIDIA TX2 high-performance computing platform, combined with the multi-sensor fusion of binocular cameras and ZED2i laser radars, and the system has the ability to operate stably in complex farmland environments. The powerful computing power of TX2 ensures the real-time performance of high-computation tasks such as deep learning models, path planning, and attitude solution, and ensures that the system can work stably under different terrains and lighting conditions. Compared with the existing technology, this system is more adaptable and can effectively cope with the variability of the picking environment.

[0045] In an exemplary embodiment, a computer device is provided, which may be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a control method for a tomato picking robot is implemented.

[0046] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0047] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0048] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0049] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0050] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0051] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0052] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A control method for a tomato picking robot, characterized in that: The control method of the tomato picking robot is applied to a tomato picking robot, and the tomato picking robot comprises: A main controller, and an environmental information acquisition module, a walking mechanism, a storage bin and a mechanical arm connected to the main controller; The environmental information acquisition module is used to obtain environmental data required for navigation and obstacle avoidance; The end of the mechanical arm is provided with an image acquisition module, a gripper and a pneumatic scissors; The image acquisition module, the clamp and the pneumatic scissors are all connected to the main controller; The control method of the tomato picking robot comprises: Acquire the environment image in front of the robot arm as the image to be recognized; The image to be identified is input into a tomato ripe fruit recognition model to identify the target in the image to be identified and the coordinates of the target in the image acquisition module coordinate system; the target is a tomato ripe fruit; the tomato ripe fruit recognition model is obtained by training a YOLO model using multiple target images; Based on the transformation matrix and the coordinates of the target in the image acquisition module coordinate system, the coordinates in the tomato picking robot base coordinate system are determined; Taking the coordinates of the tomato picking robot base coordinate system as the end point, the A* algorithm is used to plan the path of the tomato picking robot chassis to obtain the planned path; Controlling the walking mechanism to walk along the planned path, and obtaining the target posture when reaching the end point of the planned path; Moving the end of the robotic arm to a target position; Controlling the clamp and the pneumatic scissors to perform clamping and shearing operations on the target; The end of the robotic arm is controlled to clamp the target to the storage bin for storage.

2. The control method of the tomato picking robot according to claim 1, characterized in that: The image acquisition module is a binocular camera.

3. The control method of the tomato picking robot according to claim 1, characterized in that: The environmental information acquisition module includes: a depth camera and a laser radar.

4. The control method of the tomato picking robot according to claim 1, characterized in that: Before acquiring the environment image in front of the robot arm as the image to be recognized, the following steps are also included: Get multiple images containing targets; Whether the target is present and the coordinates of the target in the coordinate system of the tomato picking robot base are used as labels to annotate multiple images containing the target; With the target image as input and the label of the target image as output, the YOLO model can be trained using the loss function to obtain a tomato ripe fruit recognition model.

5. The control method of the tomato picking robot according to claim 4, characterized in that: The loss function is: ； in, is the total loss; is the position error weight coefficient; is the number of grids; is the number of bounding boxes of the i-th grid; is the indicator variable of the jth bounding box of the i-th grid; is the actual frame center coordinate; is the center coordinate of the prediction box; is the actual frame width; is the prediction box width; is the actual frame height; is the predicted box height; are other error terms.

6. The control method of the tomato picking robot according to claim 4, characterized in that: The conversion matrix includes: a conversion matrix from the image acquisition module coordinate system to the robot end coordinate system, and a conversion matrix from the robot end coordinate system to the tomato picking robot base coordinate system.

7. The control method of the tomato picking robot according to claim 1, characterized in that: Moving the end of the robotic arm to a target position, including: The end of the robot arm is moved to detect whether there is an obstacle in front of the end of the robot arm until the end of the robot arm moves to the target posture, and when an obstacle is detected, the movement of the robot arm is stopped.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the control method of the tomato picking robot according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the control method of the tomato picking robot described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the control method of the tomato picking robot described in any one of claims 1 to 7 is implemented.

Images