A robot-based safety protection method and system

Abstract

The present invention relates to the technical field of robots, and provides a safety protection method and system based on a robot. The robot is monitored by a vision camera, and a virtual vision warning line is set in the vision image of the vision camera. When an interference source enters the vision warning line, braking protection is performed to avoid collisions. In addition, the intrusion risk of the interference source is evaluated based on the vision image, and the length of the first warning distance is adjusted according to the risk assessment result. As the intrusion risk of the interference source increases, the length of the first warning distance can be increased to improve the prevention effect; when the intrusion risk of the interference source is small, the length of the first warning distance is reduced to reduce the size of the surrounding area of the vision warning line and avoid false triggering of the protection braking of the robot. The safety protection method and system based on the robot of the present invention can effectively improve the protection effect by using the virtual vision warning line, and can dynamically adjust the warning distance, taking into account the improvement of production safety and production efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of robots, and particularly to a safety protection method and system based on robots. Background Art

[0002] With the rapid development of technology, the application of robots (industrial robots, service robots, etc.) has become increasingly common. Robots basically operate along corresponding trajectories according to pre-compiled programs, but they cannot perceive and detect surrounding moving objects. During the movement of robots, there are often people or other devices operating around the robots. If the surrounding people or objects enter the working area of the robots, the robots will collide with them, causing harm to people or equipment. To avoid this problem, fences need to be installed around the robots.

[0003] However, the fences will interfere with the picking, placing, and transportation of materials. The setting of the fences has great limitations and it is difficult to form effective protection. In addition, the limbs of the human body, the extended arms of other robots or equipment, etc. are also likely to cross or pass through the fences and invade the working range of the robots, making the protection effect of physical barrier structures such as fences insufficient and the production safety insufficient. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a safety protection method and system based on robots to solve the problem of insufficient protection effect of physical barriers, resulting in insufficient production safety.

[0005] On the one hand, the present invention provides a safety protection method based on robots, including:

[0006] Collecting visual images of a visual camera set in the space above the robot, and setting a visual warning line in the visual image according to the preset working range and the first warning distance of the robot;

[0007] Judging whether there is an interference source entering the surrounded area of the visual warning line according to the visual image and the visual warning line, and when it is judged that there is an interference source entering the surrounded area of the visual warning line, controlling the robot to decelerate to a safe speed;

[0008] When the interference source exits the surrounded area of the visual warning line, controlling the robot to resume the running speed;

[0009] And, evaluating the intrusion risk of the interference source according to the visual image, and adjusting the length of the first warning distance according to the risk assessment result, wherein the higher the intrusion risk of the interference source, the longer the first warning distance.

[0010] Optionally, it further includes:

[0011] Collect the radar information of the lidar around the working range of the robot, and combine the preset working range of the robot and the second warning distance to set the radar warning area;

[0012] Judge whether there is an interference source entering the radar warning area according to the radar information and the radar warning area, and when it is judged that there is an interference source entering the radar warning area, control the robot to decelerate to a safe speed.

[0013] Optionally, the step of performing an intrusion risk assessment of the interference source according to the visual image and adjusting the length of the first warning distance according to the risk assessment result further includes:

[0014] Obtain the motion information of the interference source according to the visual image, and update the first warning distance according to the motion information, where the motion information includes the motion speed of the interference source, and the faster the motion speed of the interference source, the longer the first warning distance.

[0015] Optionally, the motion information includes the motion trajectories and motion speeds of the geometric center and geometric edge of the interference source, and the step of updating the first warning distance according to the motion information further includes:

[0016] Obtain the predicted trajectory and predicted speed of the geometric edge according to the motion trajectories and motion speeds of the geometric center and geometric edge of the interference source;

[0017] Judge whether the interference source may enter the surrounded area of the visual warning line according to the predicted trajectory of the geometric edge;

[0018] When there is a possibility that the interference source enters the surrounded area of the visual warning line, update the first warning distance according to the predicted speed of the geometric edge to update the position of the visual warning line.

[0019] Optionally, the step of performing an intrusion risk assessment of the interference source according to the visual image and adjusting the length of the first warning distance according to the risk assessment result further includes:

[0020] Perform a behavior analysis on the interference source entering the visual range according to the visual image, and obtain the current behavior danger level of the interference source in combination with a preset behavior danger level table;

[0021] Adjust the first warning distance according to the current behavior danger level, where the higher the current behavior danger level, the longer the first warning distance.

[0022] Optionally, the behaviors in the behavior danger level table include walking, running, waving, fighting, and falling. When the current behavior of the interference source includes at least two behaviors in the behavior danger level table, the current behavior danger level of the interference source is the sum of the corresponding at least two danger levels.

[0023] Optionally, it further includes: updating the first warning distance in real time according to the running speed of the robot, where the higher the running speed of the robot, the longer the first warning distance.

[0024] On the other hand, the present invention also provides a safety protection system based on a robot, including a vision camera and a main controller. The vision camera is arranged in the space above the robot, and the main controller is communicatively connected to the vision camera and the robot for:

[0025] Collecting the visual image of the vision camera arranged in the space above the robot, and setting a visual warning line in the visual image according to the preset working range and the first warning distance of the robot;

[0026] Judging whether an interference source enters the surrounded area of the visual warning line according to the visual image and the visual warning line, and controlling the robot to decelerate to a safe speed when it is judged that an interference source enters the surrounded area of the visual warning line;

[0027] When the interference source exits the surrounded area of the visual warning line, controlling the robot to resume the running speed;

[0028] And, performing an intrusion risk assessment on the interference source according to the visual image, and adjusting the length of the first warning distance according to the risk assessment result, where the higher the intrusion risk of the interference source, the longer the first warning distance.

[0029] Optionally, the main controller is further used for: obtaining the motion information of the interference source according to the visual image, and updating the first warning distance according to the motion information, where the motion information includes the motion speed of the interference source, and the faster the motion speed of the interference source, the longer the first warning distance.

[0030] Optionally, the main controller is further used for:

[0031] Performing behavior analysis on the interference source entering the visual range according to the visual image, and obtaining the current behavior danger level of the interference source in combination with a preset behavior danger level table;

[0032] Adjusting the first warning distance according to the current behavior danger level, where the higher the current behavior danger level, the longer the first warning distance.

[0033] The robot-based safety protection method provided by the present invention uses a vision camera to monitor the robot, and sets a virtual vision warning line in the vision image of the vision camera. When an interference source enters the vision warning line, braking protection is performed to avoid collisions. Among them, the intrusion risk of the interference source can also be evaluated according to the vision image, and the length of the first warning distance can be adjusted according to the risk assessment result. The higher the intrusion risk of the interference source, the longer the first warning distance. As the intrusion risk of the interference source increases, the length of the first warning distance can be increased to improve the prevention effect; when the intrusion risk of the interference source is small, the length of the first warning distance can be reduced to reduce the size of the surrounding area of the vision warning line and reduce the mis-triggering of the robot's protection braking, thereby ensuring the working efficiency of the robot. The robot-based safety protection method of the present invention can effectively avoid the setting limitations of physical fences by using a virtual vision warning line, ensure the working efficiency of the robot, can effectively monitor intrusion behaviors in all directions, improve the protection effect, and can dynamically adjust the warning distance according to the intrusion risk of the interference source. On the premise of ensuring safety, the risk of mis-triggering the protection braking and reducing production efficiency can be reduced, taking into account the improvement of production safety and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the main flowchart of the robot-based safety protection method in the embodiment of the present invention;

[0035] Figure 2 is the application installation schematic diagram of the robot-based safety protection system in the embodiment of the present invention.

[0036] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0038] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be a middle element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be a middle element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs. The terms used in the description of the invention herein are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0040] To solve the problem of insufficient protection effect of the physical barrier of the fence, resulting in insufficient production safety. The present invention provides a robot-based safety protection method, which uses a vision camera to monitor the robot, and sets a virtual vision warning line in the vision image of the vision camera. When an interference source such as a on-site worker, a transfer robot or other equipment enters the vision warning line, braking protection is carried out to avoid collisions. In addition, the intrusion risk of the interference source is evaluated according to the vision image, and the length of the first warning distance is adjusted according to the risk assessment result. As the intrusion risk of the interference source increases, the length of the first warning distance can be increased to improve the prevention effect; when the intrusion risk of the interference source is relatively small, the length of the first warning distance is reduced to reduce the size of the surrounding area of the vision warning line, and avoid accidentally triggering the protection braking of the robot and suspending production. The robot-based safety protection method and system of the present invention can effectively improve the protection effect by using a virtual vision warning line, and can dynamically adjust the warning distance, taking into account both production safety and production efficiency.

[0041] Specifically, as Figure 1 shown, is the main flowchart of the robot-based safety protection method of this embodiment, including:

[0042] Step S01: Collect the vision image of the vision camera set in the space above the robot, and set a vision warning line in the vision image according to the preset working range and the first warning distance of the robot;

[0043] Step S02: Judge whether there is an interference source entering the surrounding area of the vision warning line according to the vision image and the vision warning line, and when it is judged that there is an interference source entering the surrounding area of the vision warning line, control the robot to decelerate to a safe speed;

[0044] Step S03: When the interference source exits the surrounding area of the vision warning line, control the robot to resume the running speed;

[0045] Step S04: Evaluate the intrusion risk of the interference source according to the vision image, and adjust the length of the first warning distance according to the risk assessment result, wherein the higher the intrusion risk of the interference source, the longer the first warning distance.

[0046] In this embodiment, the vision camera is set above the robot to monitor the environment near the robot from top to bottom. Moreover, the fixed mounting base of the vision camera is separately set from the fixed mounting base of the robot to avoid the influence of the vibration during the operation of the robot on the vision camera, ensure the fixity of the visual images collected by the vision camera, and further ensure the fixity of the warning range of the set visual warning line, and ensure the warning accuracy.

[0047] In actual production, the installation position of the robot may be relatively open, or it may be against a wall, a large control cabinet or other structures. Correspondingly, the possible intrusion directions of interference sources such as external workers, other robots or equipment will be different. The lens orientation of the vision camera can be adjusted according to the possible intrusion direction of the interference source, increasing the recognition distance of the interference source, improving the protection effect, and reducing the proportion of invalid image information in the visual images, thus improving the resource utilization rate.

[0048] In step S02, the step of controlling the robot to decelerate to a safe speed is, for example, emergency braking to control the robot to stop. The corresponding ultimate stop distance is related to the first warning distance and can be obtained based on the calculation method of the robot's ultimate moving distance in the collaborative robot safety standard ISO / TS15066:2016. Generally speaking, the first warning distance is related to the speed of the interference source and the action speed of the robot. The specific calculation method is not elaborated in this application.

[0049] The safe speed is, for example, 0.25 meters per second. When a safety warning is triggered, the running speed of the robot is controlled to be reduced below 0.25 meters per second, and it runs in a relatively safe low-speed mode, which can avoid the robot from completely pausing work and reduce the energy consumption demand for restarting after shutdown.

[0050] Among them, when a safety warning is triggered, whether to control the robot to decelerate to a safe speed or stop by emergency braking can be selected according to the intrusion risk of the interference source. When the intrusion risk is relatively high and exceeds the preset risk threshold, emergency braking to stop is selected; otherwise, decelerating to a low-speed operation mode is selected.

[0051] To improve the redundancy of protection and the reliability of protection, in this embodiment, a lidar is also set outside the working area of the robot. Correspondingly, the safety protection method further includes: collecting the radar information of the lidar set around the working range of the robot, and setting a radar warning area in combination with the preset working range of the robot and the second warning distance; judging whether there is an interference source entering the radar warning area according to the radar information and the radar warning area, and controlling the robot to decelerate to a safe speed when it is judged that there is an interference source entering the radar warning area. Provide dual protection monitoring to improve the reliability of protection.

[0052] In an alternative embodiment, step S04 further includes: obtaining motion information of the interference source based on the visual image, and updating the first warning distance according to the motion information, where the motion information includes the motion speed of the interference source, and the faster the motion speed of the interference source, the longer the first warning distance.

[0053] Specifically, the motion information includes the motion trajectories and motion speeds of the geometric center and geometric edges of the interference source, and the step of updating the first warning distance according to the motion information further includes:

[0054] Obtaining the predicted trajectory and predicted speed of the geometric edge according to the motion trajectories and motion speeds of the geometric center and geometric edges of the interference source;

[0055] Judging whether the interference source may enter the surrounded area of the visual warning line according to the predicted trajectory of the geometric edge;

[0056] When there is a possibility that the interference source enters the surrounded area of the visual warning line, updating the first warning distance according to the predicted speed of the geometric edge to update the position of the visual warning line.

[0057] Taking a human body as an example, the geometric center is the human torso, and the geometric edges are the ends of the head and limbs. By superimposing the motion trajectories and motion speeds of the head and limbs with those of the human torso, the obtained predicted trajectories and predicted speeds of the head and limbs are more accurate, which can improve the recognition accuracy of security warning.

[0058] For example, when a nearby person makes a waving motion and the end of the arm moves in an arc, and it passes by near the working area of the robot but does not actually invade the working area of the robot, the motion trajectory is predicted by the above method. According to the prediction result, it can be judged that the collision risk is low, the first warning distance can be shortened, false triggering of protection can be avoided, and even if the protection is triggered at this time, the warning range is reduced, and the time taken for the intrusion source to exit the warning range is also shortened, and the time taken for the robot to resume work is shorter, thereby improving the effective working time of the robot and the production efficiency.

[0059] The motion regularity of the interference source can also be obtained based on the continuous monitoring data of the motion trajectory of the geometric edge. The confidence level is obtained based on the motion regularity. The higher the confidence level, the higher the credibility of the predicted trajectory and the predicted speed, and the higher the judgment accuracy of whether the interference source enters the surrounding area of the visual warning line; the lower the credibility, the lower the judgment accuracy of whether the interference source enters the surrounding area of the visual warning line. The predicted trajectory and the predicted speed of the geometric edge are compensated based on the confidence level. The compensated predicted trajectory is closer to the robot, and the compensated predicted speed increases. It is determined whether the interference source may enter the surrounding area of the visual warning line based on the compensated predicted trajectory and the predicted speed, and the first warning distance is updated. Among them, the lower the confidence level, the greater the compensation amount.

[0060] In production, the actions of the staff generally have obvious repeatability. When the staff performs abnormal operations, the repeatability of their actions is generally low, and when performing abnormal operations, the risk of their actions invading the working range of the robot is high. Therefore, it can be understood that the motion regularity can be confirmed according to the repeatability of their actions, and the intrusion risk can be compensated and corrected according to the motion regularity, and the first warning distance can be adjusted, which can improve the accuracy of warning protection to a certain extent and improve production safety.

[0061] In another optional embodiment, step S04 further includes:

[0062] Behavior analysis is performed on the interference source that enters the visual range according to the visual image, and the current behavior risk level of the interference source is obtained in combination with a preset behavior risk level table;

[0063] The first warning distance is adjusted according to the current behavior risk level, where the higher the current behavior risk level, the longer the first warning distance.

[0064] Among them, the behaviors in the behavior risk level table may include walking, running, waving, fighting, falling, etc. When the current behavior of the interference source includes at least two behaviors in the behavior risk level table, the current behavior risk level of the interference source is the sum of the corresponding at least two risk levels.

[0065] For example, the maximum end speed of a person waving their hand under normal circumstances is about 1.4 meters per second, while the maximum end speed during a fight is about 8 to 10 meters per second. Correspondingly, the risk levels of walking, running, waving, fighting, and falling can be set to 1, 2, 3, 5, and 4 respectively. If a person runs towards the robot and waves their hand at the same time, their risk level is 5, that is, when waving while running, the end speed of the arm will increase, and the risk of the end of the arm invading the working area of the robot and causing a collision increases, and the risk level rises. The specific level division can be continuously optimized according to historical statistical data in combination with neural network technology.

[0066] Compared with speed calculation and prediction, collision risk assessment based on human behavior can reduce the impact of calculation errors of speed and trajectory on the prediction results, making the prediction more accurate. Moreover, behavior recognition has better universality and can be achieved when the interference source is at a relatively long distance, improving the security prediction range and further enhancing the security effect.

[0067] Behavior analysis of the interference source can be realized by using neural network technology. For example, neural network models such as OpenPose, YOLO, MediaPipe, and AlphaPose can identify various actions and behaviors of the human body.

[0068] To further reduce the possibility of the robot pausing due to false triggering of protection, in this embodiment, it further includes: updating the first warning distance in real time according to the running speed of the robot, where the higher the running speed of the robot, the longer the first warning distance. Among them, when the robot transfers materials, the actions of picking up and placing materials are different from the actions of transferring materials, and their running speeds are different. The safety problem of collision mainly occurs in the stage of transferring materials. The running speed (the maximum speed during operation) of transferring materials can be used as the reference speed for adjusting the first warning distance.

[0069] To improve the accuracy of risk assessment, the current behavior of the interference source can be combined with the motion trajectory and motion speed. For example, further subdividing the danger level according to the running speed and waving speed can further improve the prediction accuracy of collision risk, improve the balance effect between safety and production efficiency, and further improve safety and production efficiency.

[0070] This application also provides a safety protection system based on a robot, as Figure 2 shown, which includes a vision camera 11 and a main controller 10. Among them, the vision camera 11 is arranged in the upper space of the robot 20 to collect the nearby environment information of the robot 20 from top to bottom. The robot 20 is, for example, a six-axis industrial robot, which can support all-directional operation requirements.

[0071] The main controller 10 is communicatively connected to the vision camera 11 and the robot 20 for safety protection according to the captured image of the vision camera 11, and controlling the robot 20 to decelerate or stop when there are safety problems. The material rack 40 is generally placed and positioned on the positioning platform 30, and the placement position of the material rack 40 is limited by the positioning platform 30, so that the material rack 40 can be accurately positioned within the effective working area of the robot 20, ensuring the working efficiency of the robot 20.

[0072] To achieve safety protection, the main controller 10 is specifically further used for:

[0073] Collect visual images of a visual camera in the space above the robot, and set a visual warning line in the visual image according to the preset working range of the robot and a first warning distance;

[0074] Judge whether there is an interference source entering the surrounded area of the visual warning line according to the visual image and the visual warning line, and when it is judged that there is an interference source entering the surrounded area of the visual warning line, control the robot to decelerate to a safe speed;

[0075] When the interference source exits from the surrounded area of the visual warning line, control the robot to resume the running speed;

[0076] And, conduct an intrusion risk assessment of the interference source according to the visual image, and adjust the length of the first warning distance according to the risk assessment result, wherein, the higher the intrusion risk of the interference source, the longer the first warning distance.

[0077] The main controller 10 is further configured to: obtain the motion information of the interference source according to the visual image, and update the first warning distance according to the motion information, wherein the motion information includes the motion speed of the interference source, and the faster the motion speed of the interference source, the longer the first warning distance.

[0078] The main controller 10 is further configured to: conduct behavior analysis on the interference source entering the visual range according to the visual image, and obtain the current behavior danger level of the interference source by combining a preset behavior danger level table; adjust the first warning distance according to the current behavior danger level, wherein, the higher the current behavior danger level, the longer the first warning distance.

[0079] The positioning platform 30 can be used for the installation and fixation of a lidar, and thus can effectively ensure the fixation of the lidar and the effectiveness of the detection result of the lidar.

[0080] The safety protection method and system based on the robot provided by the present invention can effectively improve the protection effect by using a virtual visual warning line, and can dynamically adjust the warning distance, taking into account production safety and production efficiency, reducing the possibility of the robot pausing work due to false triggering of safety protection, and thus can further improve production efficiency on the premise of ensuring safety.

[0081] And, the control of security triggering based on the warning line can isolate the software systems for risk identification and security triggering, avoiding the influence of software system failures on the effectiveness of security triggering, and thus can ensure the reliability of security triggering. When the risk identification system fails, the first warning distance can be fixed at a preset value to ensure the effectiveness of security, and the preset value of the first warning distance can be specifically selected according to specific situations.

[0082] Furthermore, the behaviors of interference sources such as people and devices can remain consistent within a relatively large time and space range, enabling the behavior recognition of interference sources to be completed at a relatively long distance, achieving risk prediction earlier. Moreover, compared with the irregular movements when the actions of interference sources are abnormal, behaviors generally have better continuity. When the interference source enters the recognition range, the behaviors generally remain consistent, and the risk prediction conclusion is relatively certain, which can ensure the accuracy of risk prediction. Therefore, based on behavior prediction, the effectiveness of risk prediction can be improved, the risk prediction distance can be effectively increased, and the security effect can be enhanced.

[0083] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0084] The above-described embodiments merely represent several specific implementation manners of the present invention, and the descriptions thereof are relatively specific and detailed. However, it should not be construed as a limitation to the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the appended claims.

Claims

1. A robot-based safety protection method, characterized in that, Including: Collecting visual images of a visual camera set in the space above the robot, and setting a visual warning line in the visual image according to a preset working range of the robot and a first warning distance; Judging whether there is an interference source entering the surrounded area of the visual warning line according to the visual image and the visual warning line, and controlling the robot to decelerate to a safe speed when it is judged that there is an interference source entering the surrounded area of the visual warning line; When the interference source exits the surrounded area of the visual warning line, controlling the robot to resume its running speed; And, performing an intrusion risk assessment of the interference source according to the visual image, and adjusting the length of the first warning distance according to the risk assessment result, wherein, the higher the intrusion risk of the interference source, the longer the first warning distance; Wherein, the step of performing an intrusion risk assessment of the interference source according to the visual image and adjusting the length of the first warning distance according to the risk assessment result further includes: Obtaining the action repeatability of the interference source according to the visual image, obtaining the motion regularity of the interference source according to the action repeatability, and compensating and correcting the intrusion risk according to the motion regularity to adjust the first warning distance, wherein, the lower the motion regularity, the greater the increment of the compensation amount for the intrusion risk, and the longer the adjusted first warning distance.

2. The robot-based safety protection method according to claim 1, wherein Further including: Collecting radar information of a lidar set around the working range of the robot, and setting a radar warning area in combination with a preset working range of the robot and a second warning distance; Judging whether there is an interference source entering the radar warning area according to the radar information and the radar warning area, and controlling the robot to decelerate to a safe speed when it is judged that there is an interference source entering the radar warning area.

3. The robot-based safety protection method according to claim 1, wherein, The step of performing an intrusion risk assessment of the interference source according to the visual image and adjusting the length of the first warning distance according to the risk assessment result further includes: Obtaining the motion information of the interference source according to the visual image, and updating the first warning distance according to the motion information, wherein, the motion information includes the motion speed of the interference source, and the faster the motion speed of the interference source, the longer the first warning distance.

4. The robot-based safety protection method according to claim 3, wherein, The motion information includes the motion trajectories and motion speeds of the geometric center and geometric edge of the interference source, and the step of updating the first warning distance according to the motion information further includes: Obtaining the predicted trajectory and predicted speed of the geometric edge according to the motion trajectories and motion speeds of the geometric center and geometric edge of the interference source; Judging whether the interference source may enter the surrounded area of the visual warning line according to the predicted trajectory of the geometric edge; When there is a possibility that the interference source enters the surrounded area of the visual warning line, updating the first warning distance according to the predicted speed of the geometric edge to update the position of the visual warning line.

5. The robot-based safety protection method according to claim 1, wherein The step of performing an intrusion risk assessment of the interference source according to the visual image and adjusting the length of the first warning distance according to the risk assessment result further includes: Perform behavioral analysis on the interference sources entering the visual range based on the visual image, and obtain the current behavioral risk level of the interference source by combining with a preset behavioral risk level table; Adjust the first warning distance according to the current behavioral risk level, wherein the higher the current behavioral risk level, the longer the first warning distance.

6. The robot-based safety protection method according to claim 5, wherein, The behaviors in the behavioral risk level table include walking, running, waving, fighting, and falling. When the current behavior of the interference source includes at least two behaviors in the behavioral risk level table, the current behavioral risk level of the interference source is the sum of the corresponding at least two risk levels.

7. The robot-based safety protection method according to claim 1, wherein It further includes: Update the first warning distance in real time according to the running speed of the robot, wherein the higher the running speed of the robot, the longer the first warning distance.

8. A robot-based safety protection system, characterized in that, It includes a visual camera and a main controller. The visual camera is arranged in the space above the robot. The main controller is communicatively connected to the visual camera and the robot for control. The main controller is used for: Collect the visual image of the visual camera arranged in the space above the robot, and set a visual warning line in the visual image according to the preset working range and the first warning distance of the robot; Judge whether there is an interference source entering the surrounded area of the visual warning line according to the visual image and the visual warning line, and when it is judged that there is an interference source entering the surrounded area of the visual warning line, control the robot to decelerate to a safe speed; When the interference source exits the surrounded area of the visual warning line, control the robot to resume the running speed; And perform an intrusion risk assessment on the interference source according to the visual image, and adjust the length of the first warning distance according to the risk assessment result. The higher the intrusion risk of the interference source, the longer the first warning distance; Wherein, the main controller is further used for: obtaining the action repeatability of the interference source according to the visual image, obtaining the motion regularity of the interference source according to the action repeatability, compensating and correcting the intrusion risk according to the motion regularity, and adjusting the first warning distance. The lower the motion regularity, the greater the increment of the compensation amount for the intrusion risk, and the longer the adjusted first warning distance.

9. The robot-based safety protection system according to claim 8, wherein, The main controller is further used for: obtaining the motion information of the interference source according to the visual image, and updating the first warning distance according to the motion information. The motion information includes the motion speed of the interference source. The faster the motion speed of the interference source, the longer the first warning distance.

10. The robot-based safety protection system according to claim 8, wherein The main controller is further used for: Perform behavioral analysis on the interference sources entering the visual range based on the visual image, and obtain the current behavioral risk level of the interference source by combining with a preset behavioral risk level table; Adjust the first warning distance according to the current behavioral risk level, wherein the higher the current behavioral risk level, the longer the first warning distance.

Images