Separated cleaning robot identification and positioning method based on binocular camera

Abstract

The invention relates to the technical field of photovoltaic cleaning vehicle three-dimensional target identification and positioning, and discloses a binocular camera-based separated cleaning robot identification and positioning method, which comprises the following steps of: assembling a photovoltaic cleaning vehicle and a separated cleaning robot, and installing a binocular camera on the cleaning vehicle; a binocular camera is connected with a camera image processing and computing unit through a network to jointly construct a robot system body, and a data acquisition module, a data computing module, a data analysis module, a result evaluation module, an execution module and a control module are established on the robot system body. The separated cleaning robot is carried on a photovoltaic sweeper, environment three-dimensional information is generated, the position of the separated cleaning robot in the three-dimensional space can be recognized in real time by combining and calculating the cleaning positioning accuracy # imgabs0 # of the robot, the relative position relation between the separated cleaning robot and the photovoltaic sweeper is determined, and therefore tracking and positioning are stably carried out; and the separation type cleaning robot can be conveniently recycled to the cleaning vehicle after avoiding obstacles.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional target recognition and positioning of photovoltaic cleaning vehicles, and specifically to a recognition and positioning method for a separable cleaning robot based on a binocular camera. Background Technique

[0002] In the technical field of positioning of separable photovoltaic cleaning robots for photovoltaic cleaning vehicles, there are currently some application methods. Among them, satellite positioning is relatively common. It requires the installation of satellite positioning equipment and the extraction of corresponding boundary information for positioning. However, this method has certain limitations. Its accuracy and stability highly depend on the strength of satellite signals. When satellite signals are blocked, interfered with, or have weak signal strength, the accuracy and reliability of positioning will be significantly affected. In addition, the method of attaching auxiliary tags to the separable robot for positioning is also applied in some scenarios. However, this method has obvious defects in the environment of dusty sand floating in the photovoltaic park. Dust is likely to accumulate on the surface of the tags, thereby reducing the accuracy of positioning and bringing many inconveniences to actual positioning operations. The defects of these existing positioning technologies have prompted the industry to further explore more effective, accurate, and environment-adaptive positioning methods to meet the high-precision positioning requirements of photovoltaic cleaning robots under different working conditions and ensure the efficient and stable development of photovoltaic cleaning operations. Summary of the Invention

[0003] (I) Technical Problems to be Solved Aiming at the deficiencies of the prior art, the present invention provides a recognition and positioning method for a separable cleaning robot based on a binocular camera, which has the advantages of high precision, high stability, and strong environmental adaptability, and solves the problems in the prior art that satellite positioning is affected by signal strength in terms of accuracy and stability, and auxiliary tag positioning is easily interfered by dust accumulation and affects accuracy.

[0004] (II) Technical Solutions To achieve the above object, the present invention provides the following technical solution: A method for identifying and positioning a separable cleaning robot based on a binocular camera, comprising the following steps: Step 1, assemble a photovoltaic cleaning vehicle and a separable cleaning robot, install the binocular camera on the cleaning vehicle, and the binocular camera is connected to the camera image processing and calculation unit through a network to jointly construct the main body of the robot system; Step 2, establish a data acquisition module, a data calculation module, a data analysis module, a result evaluation module, an execution module and a control module on the main body of the robot system; Step 3, design three units in the data acquisition module to collect data during the operation of the robot; Step 4, design three units in the data calculation module to calculate the collected data; Step 5, the data analysis module analyzes the calculation results generated in the data calculation module; Step 6, the result evaluation module evaluates the risks of collision and falling and insufficient power during the operation of the robot according to the data analysis results, and formulates corresponding warning strategies; Step 7, the execution module identifies the position of the separable cleaning robot in the three-dimensional space according to the data analysis results, determines the relative position relationship between it and the photovoltaic cleaning vehicle, generates an obstacle tracking instruction, and controls the separable cleaning robot to move along the predetermined trajectory of the system according to the optimal path planning result and the obstacle tracking instruction to complete the cleaning task; Step 8, after the separable cleaning robot completes the task, the control module guides the separable cleaning robot back to the photovoltaic cleaning vehicle and controls the separable cleaning robot to complete the docking action.

[0005] Preferably, the photovoltaic cleaning vehicle serves as the main device, responsible for recycling and placing the separable cleaning robot on the photovoltaic panel for cleaning, and the separable cleaning robot is used to perform specific cleaning tasks.

[0006] Preferably, the binocular camera is used to obtain the environmental image information of the photovoltaic cleaning vehicle, the cleaning robot and the photovoltaic panel, and position them.

[0007] Preferably, the camera image processing and calculation unit is used to process the camera image, obtain the three-dimensional information and color information of the environment, real-time locate the relative position relationship between the cleaning robot and the cleaning vehicle, and provide information to the execution module and the control module through the network for following, recycling and placing.

[0008] Preferably, the data acquisition module includes an environment perception data acquisition unit, a robot own data acquisition unit and a route obstacle data acquisition unit; the environment perception data acquisition unit acquires image data of the robot's surrounding environment in real time through a binocular camera, the robot own data acquisition unit acquires the robot's own motion data through the robot's own gyroscope, accelerometer and odometer, and the route obstacle data acquisition unit acquires the robot's route obstacle data through a picture of the required cleaning area. The environment perception data acquisition unit, the robot own data acquisition unit and the route obstacle data acquisition unit number the acquired data, and then transmit them to the data calculation module through the network for calculation.

[0009] Preferably, the robot surrounding environment image data includes point The terrain penalty and path in The robot's own motion data includes the robot's current position coordinates and the robot's own motion speed, and the robot's route obstacle data includes a set of obstacle position coordinates detected by the robot system.

[0010] Preferably, the data calculation module includes a robot obstacle avoidance unit, a robot positioning optimization unit and a robot path planning unit.

[0011] Preferably, the robot obstacle avoidance unit calculates the robot obstacle avoidance risk index based on the robot's own motion data and the robot route obstacle data. , and its calculation formula is: , in the formula, represents the robot obstacle avoidance risk index, Indicates the current position coordinates of the robot. represents the set of obstacle position coordinates detected by the robot system, and 2. ,… , Represents the total number of obstacles detected, Indicates the robot's own movement speed, Indicates the safe distance between the robot and obstacles; the robot obstacle avoidance risk index It is used to measure the risk of the robot encountering obstacles in the current motion state. The larger the value, the higher the risk.

[0012] Preferably, the robot positioning optimization unit calculates the robot cleaning positioning accuracy based on the robot's own motion data , and its calculation formula is: , in the formula, Indicates the cleaning positioning accuracy of the robot, and respectively represent the variances of the position deviations of the robot in the axis and the axis directions during cleaning, represents the maximum allowable deviation set by the robot system; the cleaning positioning accuracy of the robot is used to measure the positioning deviation of the robot during cleaning, and its value range is between When the value is closer to , it indicates higher positioning accuracy; when the value is closer to , it indicates a larger positioning deviation and lower accuracy.

[0013] Preferably, the robot path planning unit calculates the optimal motion path cost function of the robot according to the image data of the robot's surrounding environment and the robot obstacle avoidance risk index The calculation formula is: , , , in the formula, represents the optimal motion path cost function of the robot, represents the path length, that is, it represents the sum of the Euclidean distances of each segment in ={ , , …, }, represents the robot obstacle avoidance risk index, represents the environmental complexity, represents the terrain penalty of point , The turning angle of the path at is used to penalize frequent turning, , respectively represent the weight coefficients of the point terrain penalty and the path turning angle, , , respectively represent the weighting coefficients of the path length, the robot obstacle avoidance risk index and the environmental complexity in the optimal motion path cost function of the robot; the optimal motion path cost function of the robot is used to calculate the optimal motion path of the robot.

[0014] Compared with the prior art, the present invention provides a method for identifying and positioning a split cleaning robot based on a binocular camera, which has the following beneficial effects: 1. The present invention calculates the robot obstacle avoidance risk index , this formula comprehensively considers factors such as the current position of the robot, the set of obstacle positions, the robot's own movement speed, and the safety distance, and can more comprehensively reflect the obstacle risk situation faced by the robot during actual movement, making the calculated obstacle avoidance risk index more accurate. When the robot obstacle avoidance risk index has a larger value, it indicates a higher risk at this time. In the case of multiple obstacles, it will automatically trigger an alarm prompt of the robot system. Through the processing of the obstacle position coordinate set by relevant staff, it can effectively handle much more complex multi-obstacle scenarios and provide strong support for the path planning and movement speed control of the robot in a multi-obstacle environment. 2. By adopting the vision technology of a binocular camera in the present invention and mounting it on a photovoltaic cleaning vehicle to generate three-dimensional environmental information, and then combining it with the computer robot cleaning positioning accuracy , it can real-time identify the position of the separable cleaning robot in three-dimensional space, determine its relative position relationship with the photovoltaic cleaning vehicle, so as to stably perform tracking and positioning, facilitate the separable cleaning robot to be recovered onto the cleaning vehicle, and does not require additional information collection, with stable and reliable performance, solving the problems in the prior art that satellite positioning is affected by signal strength in terms of accuracy and stability, and auxiliary tag positioning is easily interfered by dust accumulation in terms of accuracy. 3. By calculating the optimal motion path cost function of the robot , and then generating an obstacle tracking instruction according to the robot obstacle avoidance risk index , from the calculation formula of the optimal motion path cost function , the optimal path planning result is obtained. By comprehensively considering the optimal path planning result and the obstacle tracking instruction, the separable cleaning robot is controlled to move along the system's predetermined trajectory to complete the cleaning task. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] Please refer to Figure 1, A recognition and positioning method for a separable cleaning robot based on a binocular camera, comprising the following steps: Step 1, Assemble a photovoltaic cleaning vehicle and a separable cleaning robot, install the binocular camera on the cleaning vehicle, and connect the binocular camera to the camera image processing and calculation unit through a network to jointly construct the main body of the robot system; Step 2, Establish a data acquisition module, a data calculation module, a data analysis module, a result evaluation module, an execution module, and a control module on the main body of the robot system to ensure that the parameters of each module are set correctly; Step 3, Design three units in the data acquisition module to collect data during the operation of the robot; Step 4, Design three units in the data calculation module to calculate the collected data; Step 5, The data analysis module analyzes the calculation results generated in the data calculation module, specifically: analyze the generated three-dimensional information and environmental parameter data, identify obstacles and photovoltaic panel distribution information in the environment, analyze the results of the robot positioning optimization unit, determine the current precise position and posture of the robot, compare with the preset position, calculate the deviation, and analyze the feasibility and safety of the path according to the path generated by the robot path planning unit and in combination with real-time environmental data; Step 6, The result evaluation module evaluates the risks of collision and falling and insufficient power during the operation of the robot according to the data analysis results and formulates corresponding warning strategies; Step 7, The execution module identifies the position of the separable cleaning robot in the three-dimensional space according to the data analysis results, determines the relative position relationship between it and the photovoltaic cleaning vehicle, generates an obstacle tracking instruction, and controls the separable cleaning robot to move along the predetermined trajectory of the system according to the optimal path planning result and the obstacle tracking instruction to complete the cleaning task; Step 8, After the separable cleaning robot completes the task, the control module guides the separable cleaning robot back to the photovoltaic cleaning vehicle and controls the separable cleaning robot to complete the docking action.

[0018] The photovoltaic cleaning vehicle, as the main device, is responsible for recycling and deploying the separable cleaning robot onto the photovoltaic panel for cleaning, and the separable cleaning robot is used to perform specific cleaning tasks.

[0019] The binocular camera is used to obtain the environmental image information of the photovoltaic cleaning vehicle, the cleaning robot, and the photovoltaic panel and position them.

[0020] The advantages are: By adopting the vision technology of the binocular camera, mounting it on the photovoltaic cleaning vehicle to generate three-dimensional environmental information, and then combining with calculating the cleaning positioning accuracy of the robot , it can identify the position of the separable cleaning robot in the three-dimensional space in real time, determine the relative position relationship between it and the photovoltaic cleaning vehicle, so as to stably perform tracking and positioning, facilitate the separable cleaning robot to be recovered onto the cleaning vehicle, and does not require additional information collection, with stable and reliable performance, solving the problems that satellite positioning in the prior art is affected by signal strength in terms of accuracy and stability, and auxiliary tag positioning is easily interfered by dust accumulation in terms of accuracy.

[0021] The camera image processing calculation unit is used to process the camera image, obtain the three-dimensional information and color information of the environment, locate the relative position relationship between the cleaning robot and the sweeper in real time, and provide information to the execution module and control module through the network for following, recovery and delivery.

[0022] The data acquisition module includes an environmental perception data acquisition unit, a robot's own data acquisition unit and a route obstacle data acquisition unit; the environmental perception data acquisition unit collects image data of the robot's surrounding environment in real time through a binocular camera, providing basic data for environmental perception and three-dimensional information generation; the robot's own data acquisition unit collects the robot's own motion data through the robot's built-in gyroscope, accelerometer and odometer for subsequent positioning optimization calculations; the route obstacle data acquisition unit obtains the robot's route obstacle data through the image of the required cleaning area, providing a basis for path planning; the environmental perception data acquisition unit, the robot's own data acquisition unit and the route obstacle data acquisition unit number the collected data, and then transmit it to the data calculation module through the network for calculation.

[0023] The robot's surrounding image data includes point The terrain penalty and path in The robot's own motion data includes the robot's current position coordinates and the robot's own motion speed; the robot's route obstacle data includes the set of obstacle position coordinates detected by the robot system.

[0024] The data calculation module includes a robot obstacle avoidance unit, a robot positioning optimization unit and a robot path planning unit.

[0025] The robot obstacle avoidance unit calculates the robot obstacle avoidance risk index based on the robot's own motion data and robot route obstacle data , and its calculation formula is: , in the formula, represents the robot obstacle avoidance risk index, Indicates the current position coordinates of the robot. represents the set of obstacle position coordinates detected by the robot system, and 2. ,… , Represents the total number of obstacles detected, Indicates the robot's own movement speed, Indicates the safe distance between the robot and obstacles; robot obstacle avoidance risk index It is used to measure the risk of the robot encountering obstacles in the current motion state. The larger the value, the higher the risk.

[0026] The advantage is that the robot obstacle avoidance risk index The calculation method comprehensively considers the current position of the robot, the position set of obstacles, the robot's own movement speed, and the factor of safety distance, and can more comprehensively reflect the obstacle risk situation faced by the robot during actual movement, making the calculated obstacle avoidance risk index more accurate. When the robot obstacle avoidance risk index has a larger value, it indicates that the risk is higher at this time, and there will be a situation of multiple obstacles, which will automatically trigger an alarm prompt of the robot system. Through the processing of the obstacle position coordinate set by relevant staff, it can effectively cope with much more complex obstacle scenarios, providing strong support for the path planning and movement speed control of the robot in a multi-obstacle environment.

[0027] The robot positioning optimization unit calculates the cleaning positioning accuracy of the robot according to the robot's own movement data , and its calculation formula is: , in the formula, represents the cleaning positioning accuracy of the robot, and respectively represent the variances of the position deviations of the robot in the axis and axis directions during cleaning, represents the maximum allowable deviation set by the robot system; The cleaning positioning accuracy of the robot is used to measure the positioning deviation of the robot during cleaning, and its value range is between . The closer the value is to , the higher the positioning accuracy; the closer the value is to , the larger the positioning deviation and the lower the accuracy.

[0028] The advantage is that by calculating the cleaning positioning accuracy of the robot , the position of the split cleaning robot in three-dimensional space is identified, and its relative position relationship with the photovoltaic cleaning vehicle is determined, providing a clear direction basis for the optimization of the robot positioning system. When it is monitored that the cleaning positioning accuracy of the robot is close to , it indicates that the positioning deviation is larger and the accuracy is lower. The system will specifically analyze the deviation situations in the axis and axis directions, automatically find the reasons for the excessive deviation, such as sensor accuracy problems and algorithm defects, and then take corresponding improvement measures to gradually improve the cleaning positioning accuracy of the robot.

[0029] The robot path planning unit calculates the cost function of the optimal motion path of the robot based on the image data of the environment around the robot and the robot obstacle avoidance risk index to calculate the cost function of the optimal motion path of the robot , and its calculation formula is: , , in the formula, represents the cost function of the optimal motion path of the robot, represents the path length, that is, it represents ={ , , …, } is the sum of the Euclidean distances of each segment, represents the robot obstacle avoidance risk index, represents the environmental complexity, represents the terrain penalty of point , the turning angle of the path at is used to penalize frequent turning, , respectively represent the weight coefficients of the point terrain penalty and the path turning angle, , , respectively represent the weighted coefficients of the path length, the robot obstacle avoidance risk index and the environmental complexity in the cost function of the optimal motion path of the robot; the cost function of the optimal motion path of the robot is used to calculate the optimal motion path of the robot

[0030] The advantages are: by calculating the cost function of the optimal motion path of the robot , generating obstacle tracking instructions according to the robot obstacle avoidance risk index , obtaining the optimal path planning result according to the cost function of the optimal motion path , and controlling the separate cleaning robot to move along the system predetermined trajectory to complete the cleaning task by integrating the optimal path planning result and the obstacle tracking instructions

[0031] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents

Claims

1. A method for identifying and locating a separate cleaning robot based on a binocular camera, characterized in that: The following steps are involved: Step 1: Assemble the photovoltaic sweeper and the detachable cleaning robot, install the binocular camera on the sweeper, connect the binocular camera to the camera image processing and computing unit through the network, and jointly build the robot system body; Step 2: Establish the data acquisition module, data computing module, data analysis module, result evaluation module, execution module and control module on the robot system body; Step 3: Design three units in the data acquisition module to collect data during the robot's working process; Step 4: Design three units in the data computing module to calculate the collected data; Step 5: The data analysis module analyzes the calculation results generated in the data computing module; Step 6: Conclusion The result evaluation module evaluates the risk of collision and falling and the risk of insufficient power during the robot's operation according to the data analysis results, and formulates corresponding early warning strategies; Step seven, the execution module identifies the position of the detachable cleaning robot in three-dimensional space according to the data analysis results, and determines the relative position relationship between it and the photovoltaic sweeper, generates obstacle tracking instructions, and controls the detachable cleaning robot to move according to the system's predetermined trajectory according to the optimal path planning results and the obstacle tracking instructions to complete the cleaning task; Step eight, after the detachable cleaning robot completes the task, the control module guides the detachable cleaning robot to return to the photovoltaic sweeper, and controls the detachable cleaning robot to complete the docking action.

2. According to claim 1, a binocular camera-based separation cleaning robot identification and positioning method is characterized in that: The photovoltaic cleaning vehicle is a main device responsible for collecting and placing the detachable cleaning robot onto the photovoltaic panel for cleaning. The detachable cleaning robot is used to perform specific cleaning tasks.

3. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 1, characterized in that: The binocular camera is used to obtain environmental image information of the photovoltaic sweeper, cleaning robot, and photovoltaic panel, and to locate them.

4. The method for identifying and locating a separate cleaning robot based on a binocular camera according to claim 1, characterized in that: The camera image processing calculation unit is used to process the camera image, obtain the three-dimensional information and color information of the environment, locate the relative position relationship between the cleaning robot and the sweeper in real time, and provide information to the execution module and the control module through the network for following, recovery and delivery.

5. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 1, characterized in that: The data acquisition module includes an environment perception data acquisition unit, a robot own data acquisition unit and a route obstacle data acquisition unit; the environment perception data acquisition unit acquires image data of the robot's surrounding environment in real time through a binocular camera, the robot own data acquisition unit acquires the robot's own motion data through the robot's own gyroscope, accelerometer and odometer, and the route obstacle data acquisition unit acquires the robot's route obstacle data through a picture of the required cleaning area. The environment perception data acquisition unit, the robot own data acquisition unit and the route obstacle data acquisition unit number the acquired data and then transmit them to the data calculation module through the network for calculation.

6. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 5, characterized in that: The robot surrounding environment image data includes points The terrain penalty and path in The robot's own motion data includes the robot's current position coordinates and the robot's own motion speed, and the robot's route obstacle data includes a set of obstacle position coordinates detected by the robot system.

7. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 1, characterized in that: The data calculation module includes a robot obstacle avoidance unit, a robot positioning optimization unit and a robot path planning unit.

8. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 7, characterized in that: The robot obstacle avoidance unit calculates the robot obstacle avoidance risk index based on the robot's own motion data and the robot route obstacle data. , and its calculation formula is: , in the formula, represents the robot obstacle avoidance risk index, Indicates the current position coordinates of the robot. represents the set of obstacle position coordinates detected by the robot system, and 2. ,… , Represents the total number of obstacles detected, Indicates the robot's own movement speed, Indicates the safe distance between the robot and obstacles; the robot obstacle avoidance risk index It is used to measure the risk of the robot encountering obstacles in the current motion state. The larger the value, the higher the risk.

9. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 7, characterized in that: The robot positioning optimization unit calculates the robot cleaning positioning accuracy based on the robot's own motion data , and its calculation formula is: , in the formula, Indicates the robot cleaning positioning accuracy. and They represent the robot's cleaning process. Axis and The variance of the position deviation in the axis direction, Indicates the maximum allowable deviation set by the robot system; The robot cleaning positioning accuracy It is used to measure the positioning deviation of the robot during cleaning, and its value range is The closer the value is, When the value is close to , it means the larger the positioning deviation is, the lower the accuracy is.

10. The method for identifying and locating a detachable cleaning robot based on a binocular camera according to claim 7, characterized in that: The robot path planning unit calculates the robot path based on the robot's surrounding environment image data and the robot's obstacle avoidance risk index. Calculate the cost function of the robot's optimal motion path , and its calculation formula is: , , in the formula, represents the cost function of the robot's optimal motion path, represents the path length, that is, ={ , , …, The sum of the Euclidean distances of each segment in}, represents the robot obstacle avoidance risk index, Indicates the complexity of the environment, Indicate point The terrain penalty, The path is The turning angle at is used to punish frequent turns. , Respectively represent the weight coefficients of point terrain penalty and path turning angle, , , They represent the weighted coefficients of path length, robot obstacle avoidance risk index and environment complexity in the robot's optimal motion path cost function; The robot's optimal motion path cost function Used to calculate the optimal motion path of the robot.

Images