Sky rail robot obstacle recognition method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of obstacle recognition, and provides a sky rail robot obstacle recognition method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the obstacle recognition based on laser radar point cloud data, and determining an environment model in a world coordinate system, and the world coordinate system coordinates of obstacles around a sky rail robot; determining target area data based on an environment model and the world coordinate system coordinates of the obstacle, and performing sliding window traversal on target area features corresponding to the target area data to obtain a first convolution calculation result of data in a window; and determining a first obstacle recognition result of the sky rail robot based on the first convolution calculation result. In the process, an obstacle recognition result is determined based on a convolution calculation result, so that whether an obstacle exists in the operation target area of the sky rail robot or not can be efficiently recognized, the risk of safety accidents is further reduced, and the safety of the production process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of obstacle recognition, and in particular to an obstacle recognition method, device, electronic equipment and storage medium for an overhead rail robot. Background Art

[0002] In the refined production process of liquor, intelligent overhead rail robots play a vital role, as they are required to automatically and accurately grab the dregs from the cellar.

[0003] However, given the unique process requirements of liquor brewing, some production sites still inevitably require workers to directly enter the cellar to perform operations. In this case, if the overhead robot cannot accurately judge the real-time conditions of the cellar area and the inside of the cellar, such as when avoiding parts passing over the workers' heads, people or objects in the cellar continue to enter and exit the cellar, the risk of safety accidents will be greatly increased. Summary of the invention

[0004] The present invention provides an overhead rail robot obstacle recognition method, device, electronic equipment and storage medium, which are used to solve the defect in the prior art that the overhead rail robot cannot accurately judge the real-time conditions of the cellar area and the inside of the cellar, thereby greatly increasing the risk of safety accidents.

[0005] The present invention provides an obstacle recognition method for an overhead rail robot, comprising the following steps: Obstacle identification is performed based on the LiDAR point cloud data to determine the environment model in the world coordinate system and the world coordinate system coordinates of the obstacles around the Tiangan robot; Based on the environment model and the coordinates of the obstacle world coordinate system, determine the target area data, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window; Based on the first convolution calculation result, a first obstacle recognition result of the overhead rail robot is determined.

[0006] According to an obstacle recognition method for an overhead rail robot provided by the present invention, the formula of the environment model is as follows: in, , is the origin of the radar device in world coordinates obtained from the controller. To adjust the sign of the direction, is the radar laser ranging value, is the horizontal angle of the radar ray, is the vertical angle of the radar, is the vertical rotation angle of the motor, The coordinates of the data points in world coordinates obtained by aligning the converted radar scan.

[0007] According to an overhead rail robot obstacle recognition method provided by the present invention, determining a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result includes: Determining a result score based on the first convolution calculation result, and determining that the first obstacle recognition result is that an obstacle exists when the result score is greater than a threshold score; The threshold score is obtained through training based on first sample data with obstacles, second sample data without obstacles, and label results.

[0008] According to an overhead rail robot obstacle recognition method provided by the present invention, the target area data is determined based on the environment model and the coordinates of the obstacle world coordinate system, including: The working process of the Skyrail robot includes a moving process; When the overhead rail robot is in the moving process, the target area data is determined based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot, and the target area data is the alarm area data; the first obstacle coordinates and the second obstacle coordinates are the obstacle world coordinate system coordinates obtained by two consecutive scans, respectively.

[0009] According to an overhead rail robot obstacle recognition method provided by the present invention, the target area data is determined based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot, and then further includes: When the second obstacle recognition result corresponding to the target area data indicates that an obstacle exists, controlling the overhead rail robot to decelerate until it stops, and issuing a warning signal; The relative position of the obstacle and the avoidance component of the overhead rail robot is determined, and the avoidance component is adjusted based on the relative position.

[0010] According to an overhead rail robot obstacle identification method provided by the present invention, the working process of the overhead rail robot also includes a process of entering and exiting a pool; Using the cellar area data as the alarm area data; Performing a sliding window traversal on the alarm area features corresponding to the alarm area data to obtain a third convolution calculation result of the data in the window; Based on the third convolution calculation result, a third obstacle recognition result of the overhead rail robot is determined, and when the third obstacle recognition result indicates that an obstacle exists, an alarm signal is issued.

[0011] According to an overhead rail robot obstacle recognition method provided by the present invention, the target area data is determined based on the environment model and the coordinates of the obstacle world coordinate system, including: Determining first target area data based on the environment model and the obstacle world coordinate system coordinates; Filling the missing data or abnormal data in the first target area data to obtain second target area data; The second target area data is sampled by using one of group sampling and cluster sampling to obtain the target area data.

[0012] The present invention also provides an obstacle recognition device for an overhead rail robot, comprising the following modules: A determination unit, used to identify obstacles based on the laser radar point cloud data, determine the environment model in the world coordinate system, and the world coordinate system coordinates of obstacles around the Tiangan robot; A target area data determination unit is used to determine the target area data based on the environment model and the coordinates of the obstacle world coordinate system, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window; The obstacle recognition result determination unit is used to determine a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an obstacle recognition method for an overhead rail robot as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the obstacle recognition method for the overhead rail robot as described in any one of the above-mentioned methods is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the obstacle recognition method for the overhead rail robot as described above is implemented.

[0016] The obstacle identification method, device, electronic device and storage medium of the overhead rail robot provided by the present invention perform obstacle identification based on laser radar point cloud data, determine the environment model in the world coordinate system, and the world coordinate system coordinates of the obstacles around the overhead rail robot; determine the target area data based on the environment model and the world coordinate system coordinates of the obstacles, perform sliding window traversal on the target area features corresponding to the target area data, and obtain the first convolution calculation result of the data in the window; determine the first obstacle identification result of the overhead rail robot based on the first convolution calculation result. This process determines the obstacle identification result based on the convolution calculation result, so that it can efficiently identify whether there are obstacles in the target area of ​​the overhead rail robot operation, further reduce the risk of safety accidents, and improve the safety of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of the obstacle recognition method of the overhead rail robot provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the installation position of the 2D scanner component provided by the present invention.

[0020] Figure 3 It is a schematic diagram of obstacle avoidance during the traveling condition of the overhead rail robot provided by the present invention.

[0021] Figure 4 It is a schematic diagram of the avoidance of the traveling working condition of the overhead rail robot provided by the present invention.

[0022] Figure 5 It is a structural schematic diagram of the obstacle recognition device of the overhead rail robot provided by the present invention.

[0023] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, rather than to describe a specific order or precedence. It should be understood that the terms used in this way can be interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type.

[0026] Figure 1 is a flow chart of the obstacle recognition method of the overhead rail robot provided by the present invention, such as Figure 1 As shown, the method includes step 110 , step 120 and step 130 .

[0027] Step 110, perform obstacle recognition based on the laser radar point cloud data, determine the environment model in the world coordinate system, and the world coordinate system coordinates of the obstacles around the Tiangan robot.

[0028] Specifically, Figure 2 is a schematic diagram of the installation position of the 2D scanner component provided by the present invention, such as Figure 2 As shown, the transfer assembly end of the Tiangan robot has a good detection field of view whether it is moving or entering and exiting the pool. Installing the 2D scanner component at the transfer assembly end facilitates timely and clear scanning and detection of the target area.

[0029] Here, the 2D scanner components include laser radar and its supporting equipment, rotating motor, reducer, IMU (Inertial Measurement Unit) sensor and other equipment. The rotating motor can help the radar obtain all-round scanning data, and the reducer can ensure that the origin is not lost during the rotation process, thereby ensuring the stability of the radar data.

[0030] The laser radar can be a multi-line laser radar, or a combination of multiple single-line laser radars, which is not specifically limited in the embodiment of the present invention. The laser radar of the 2D scanner assembly is used to scan the surrounding environment of the sky rail robot to obtain laser radar point cloud data.

[0031] Among them, the LiDAR point cloud data contains the three-dimensional coordinate information of each point in the three-dimensional space. According to the LiDAR point cloud data, obstacles can be identified, and the point cloud data belonging to the same object can be aggregated to obtain the obstacles around the Tianrail robot. In addition, according to the LiDAR point cloud data, the environmental model in the world coordinate system can be determined, and the length, width and height of the obstacle can be calculated, as well as the world coordinate system coordinates of the obstacle, and the relative position information between each obstacle and the Tianrail robot. In addition, obstacles can be tracked and each obstacle target can be assigned a unique tracking ID to identify each obstacle.

[0032] An inertial measurement unit is a device used to measure and report the acceleration and rotation rate of an object in a specific reference frame. An IMU usually consists of three accelerometers and three gyroscopes, which are used to measure the linear acceleration and angular velocity of an object in three axes (usually X, Y, and Z axes). These data can be used to calculate information such as the position, direction, and speed of an object.

[0033] Step 120, based on the environment model and the coordinates of the obstacle world coordinate system, determine the target area data, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window.

[0034] Specifically, when obstacle avoidance is turned on, the laser radar continuously scans and obtains all-round point cloud data of the environment as the motor rotates. The controller is used to align and integrate the radar data, the motor rotation angle, and the three-dimensional coordinates of the robot, so that the point cloud data can be used to restore and obtain the environment model in world coordinates. Even if the robot is in motion, the method can obtain the environment model without being affected because the controller can obtain the radar origin in world coordinates.

[0035] Among them, the formula of the environment model in the world coordinate system is as follows: in, , is the origin of the radar device in world coordinates obtained from the controller. To adjust the sign of the direction, is the radar laser ranging value, is the horizontal angle of the radar ray, is the vertical angle of the radar, is the vertical rotation angle of the motor, The coordinates of the data points in world coordinates obtained by aligning the converted radar scan.

[0036] After the environment model and the coordinates of the obstacle world coordinate system are obtained, the target area data may be determined based on the environment model and the coordinates of the obstacle world coordinate system.

[0037] That is, the world coordinate system coordinates of the obstacles are filtered out from the environment model to obtain the target area data.

[0038] It is understandable that it is also possible to formulate appropriate rules in combination with the action logic of the overhead rail robot and the process flow of on-site production, and determine the target area data based on the appropriate rules and environmental models.

[0039] In addition, considering the rectilinear propagation characteristics of radar light, data loss may occur due to high obstruction. In addition, floating objects in the air or radar hardware problems may cause outlier noise data points in the data. These missing data points and outlier noise points will affect the result analysis. Missing values ​​and outlier value processing can be performed on the target area data.

[0040] After the target area data is determined, a sliding window traversal is performed on the target area features corresponding to the target area data to obtain a first convolution calculation result of the data in the window.

[0041] It is understandable that sliding window traversal can be implemented by Convolutional Neural Network (CNN), which is highly robust to common interference factors such as illumination changes, image blur, occlusion, etc. This means that even under adverse conditions, the Tiantui robot can maintain stable obstacle recognition performance, ensuring safe and reliable operation.

[0042] In actual use, the convolution kernel is used for the target area data obtained in real time. , for the target area features Perform sliding window traversal to obtain the data in the window The first convolution calculation result is as follows: in, represents the convolution kernel, Indicates the number of rows, Indicates the number of columns, Respectively represent the horizontal index and vertical index corresponding to the current window. Represents the target area characteristics.

[0043] Step 130: Determine a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0044] Specifically, after obtaining the first convolution calculation result, a result score can be determined based on the first convolution calculation result, and a first obstacle recognition result of the overhead rail robot can be determined based on the result score and the threshold score.

[0045] It should be noted that once a person or foreign object is detected, the robot will immediately pause or stop its automatic operation and trigger the warning and alarm system to alert relevant personnel to potential safety hazards. After detecting safety or clearing the alarm signal, the robot will continue its automatic operation process to ensure the safety and smoothness of the entire production process.

[0046] The method provided by the embodiment of the present invention performs obstacle recognition based on lidar point cloud data to determine the environment model in the world coordinate system and the world coordinate system coordinates of the obstacles around the overhead rail robot; based on the environment model and the world coordinate system coordinates of the obstacles, the target area data is determined, and a sliding window traversal is performed on the target area features corresponding to the target area data to obtain a first convolution calculation result of the data in the window; based on the first convolution calculation result, a first obstacle recognition result of the overhead rail robot is determined, so that it can efficiently identify whether there are obstacles in the target area of ​​the overhead rail robot's operation, further reduce the risk of safety accidents, and improve the safety of the production process.

[0047] Based on the above embodiment, step 130 includes: Step 131: determining a result score based on the first convolution calculation result, and determining that the first obstacle recognition result is that an obstacle exists when the result score is greater than a threshold score; The threshold score is obtained through training based on first sample data with obstacles, second sample data without obstacles, and label results.

[0048] Specifically, based on the first convolution calculation result, the result score is determined , and in the resulting score Greater than threshold score In the case where the first obstacle recognition result is determined to be the existence of an obstacle, an alarm signal is sent to the device.

[0049] Here, the threshold score can be obtained based on the following steps: First, first sample data with obstacles, second sample data without obstacles, and label results are collected in advance. The label results usually indicate whether the sample data contains obstacles, for example, 1 indicates that obstacles are contained, and 0 indicates that obstacles are not contained.

[0050] Then, the first sample data and the second sample data are cleaned, and a convolution kernel is generated based on the first sample data and the second sample data after data cleaning. and threshold score .

[0051] Specifically, for binary classification problems (such as determining whether there is an obstacle), commonly used models include logistic regression, support vector machines, convolutional neural networks, etc.

[0052] During model training, one or more threshold scores are usually set to convert the model output into the final classification result. For logistic regression models, the threshold score is usually set to 0.5 (i.e., samples with a probability greater than 0.5 are classified as positive, and samples with a probability less than 0.5 are classified as negative). For other models, the setting of the threshold score may need to be adjusted according to the actual situation, and in the process, the convolution kernel of the training process is retained. .

[0053] Based on the above embodiment, determining the target area data based on the environment model and the obstacle world coordinate system coordinates in step 120 includes: Step 121, when the overhead rail robot is in the moving process, based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot, determine the target area data, the target area data is the alarm area data; the first obstacle coordinates and the second obstacle coordinates are the obstacle world coordinate system coordinates obtained by two consecutive scans respectively.

[0054] Specifically, Figure 3 Schematic diagram of the obstacle avoidance of the overhead rail robot provided by the present invention. Figure 3 As shown, the working process of the overhead rail robot includes a moving process. When the overhead rail robot is in the moving process, the target area data can be determined based on the environmental model, the first obstacle coordinates and the second obstacle coordinates, as well as the moving speed and moving direction of the overhead rail robot. Here, the target area data is the alarm area data, and then based on the alarm area data, it is determined whether an obstacle will appear in the alarm area.

[0055] Here, the first obstacle coordinate and the second obstacle coordinate are the coordinates of the obstacle world coordinate system obtained by two consecutive scans. The moving speed and moving direction of the overhead rail robot are obtained by the controller.

[0056] It is understandable that by obtaining the first obstacle coordinates and the second obstacle coordinates, the specific location of the obstacle can be clearly known. These coordinate data provide the robot with accurate information about the location of the obstacle, which is the basis for determining the alarm area. Combined with the environmental model, the alarm area can be accurately determined.

[0057] By knowing the moving speed and direction of the Skyrail robot, we can predict the robot's moving path in the future. By combining the predicted path with the obstacle coordinates, we can determine whether the robot has a potential collision risk with the obstacle, that is, we can determine the alarm area.

[0058] Then, a sliding window traversal is performed on the alarm area features corresponding to the alarm area data to obtain a second convolution calculation result of the data in the window, and then based on the second convolution calculation result, a second obstacle recognition result of the overhead rail robot is determined.

[0059] It can be understood that when the second obstacle recognition result corresponding to the target area data is that there is an obstacle, the overhead rail robot is controlled to slow down until it stops and sends out a warning signal.

[0060] Figure 4 Schematic diagram of the avoidance of the overhead rail robot provided by the present invention. Figure 4 As shown, while sending out an alarm signal, the overhead rail robot will continue to scan, determine the relative position of the obstacle and the avoidance component of the overhead rail robot, and adjust the avoidance component based on the relative position.

[0061] Here, the avoidance component, such as the bucket of the overhead rail robot, can be adjusted left and right until the obstacle leaves the alarm range and the overhead rail robot continues to move. If it cannot be avoided, the warning signal will be upgraded to an alarm signal after a certain period of time, prompting the management personnel to intervene.

[0062] In addition, during the operation of the Tiangan robot, severe vibration may cause abnormal mutations in the radar origin coordinates. The Tiangan robot controller can judge the IMU sensor signal on the radar device in real time. When the vibration exceeds the preset threshold and conditions, the radar scanning data is judged to be abnormal. The controller shields the abnormal part of the radar data to avoid interference with the detection results.

[0063] Based on the above embodiment, the working process of the overhead rail robot also includes the process of entering and exiting the pool; Using the cellar area data as the alarm area data; Performing a sliding window traversal on the alarm area features corresponding to the alarm area data to obtain a third convolution calculation result of the data in the window; Based on the third convolution calculation result, a third obstacle recognition result of the overhead rail robot is determined, and when the third obstacle recognition result indicates that an obstacle exists, an alarm signal is issued.

[0064] Specifically, when the robot is in the process of entering or exiting the pool, the robot is stationary above the cellar area, and the radar rotates to collect data for judgment. The cellar area data is used as the alarm area data, and then the alarm area features corresponding to the alarm area data are traversed by sliding windows to obtain the third convolution calculation result of the data in the window.

[0065] Based on the third convolution calculation result, a third obstacle recognition result of the overhead rail robot is determined, and when the third obstacle recognition result indicates that an obstacle exists, an alarm signal is issued.

[0066] It should be noted that as long as an obstacle is detected inside the cellar, the overhead rail robot will stop moving and send out an alarm signal. Due to the depth difference of the cellar, workers need to intervene to eliminate abnormal situations. After the workers have dealt with the abnormalities inside the cellar, the alarm is cleared and the overhead rail robot continues to move.

[0067] Based on the above embodiment, the step 120 of determining the target area data based on the environment model and the obstacle world coordinate system coordinates includes: Step 210, determining first target area data based on the environment model and the obstacle world coordinate system coordinates; Step 220, filling the missing data or abnormal data in the first target area data to obtain second target area data; Step 230: Sampling the second target area data by using one of group sampling and cluster sampling to obtain the target area data.

[0068] Specifically, the data of the first target area can be determined based on the environmental model and the coordinates of the obstacle world coordinate system. Considering that for the acquired radar data of the target area, due to the rectilinear propagation characteristics of the radar light, there will be data missing due to high obstruction. In addition, floating objects in the air or radar hardware problems will cause outlier noise data points in the data. These missing data points and outlier noise points will affect the result analysis.

[0069] Then, the missing data or abnormal data in the first target area data is filled to obtain the second target area data.

[0070] It is understandable that methods other than K-means can be used to fill in missing data and remove abnormal noise data.

[0071] The formula for filling the missing data or abnormal data in the first target area data is as follows: in, is a missing point or an abnormal point, The closest training data points, are the actual values ​​corresponding to these points, yes and The distance between It is The weight of the nearest neighbor is Inversely proportional, p Is a positive real number used to adjust the influence of distance on weight.

[0072] In order to appropriately reduce the demand for hardware computing performance at the front-line production site while retaining environmental characteristics, one of group sampling and cluster sampling can be used to sample the second target area data to obtain the target area data.

[0073] Based on any of the above embodiments, in an embodiment of the present invention, first, the point cloud data in the target area is collected by installing a laser radar device at the end of the transfer unit of the overhead rail robot. Then, the collected data is effectively cleaned and sampled to obtain the characteristic information in the target scanning area; then, using efficient algorithms such as sliding windows and convolution calculations, this information is accurately analyzed to calculate the possibility of obstacles in the target area; finally, when the calculation result exceeds the preset warning threshold or reaches the preset alarm condition, the system determines that there are obstacles in the target area and performs subsequent obstacle avoidance processes in a targeted manner. The application of this method significantly improves the safety and reliability of the liquor production process, effectively reduces potential safety risks, and can continue to be optimized during use.

[0074] The overhead rail robot obstacle recognition device provided by the present invention is described below. The overhead rail robot obstacle recognition device described below and the overhead rail robot obstacle recognition method described above can be referred to each other.

[0075] Based on any of the above embodiments, the present invention provides an obstacle recognition device for an overhead rail robot. Figure 5 Schematic diagram of the structure of the obstacle recognition device of the overhead rail robot provided by the present invention. Figure 5 As shown, the device comprises: A determination unit 510 is used to identify obstacles based on the laser radar point cloud data, determine the environment model in the world coordinate system, and the world coordinate system coordinates of obstacles around the Tiangan robot; A target area data determination unit 520 is used to determine the target area data based on the environment model and the obstacle world coordinate system coordinates, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window; The obstacle recognition result determination unit 530 is used to determine a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0076] The device provided by the embodiment of the present invention performs obstacle recognition based on lidar point cloud data, determines the environment model in the world coordinate system, and the world coordinate system coordinates of the obstacles around the overhead rail robot, and determines the target area data based on the environment model and the world coordinate system coordinates of the obstacles, performs a sliding window traversal on the target area features corresponding to the target area data, and obtains a first convolution calculation result of the data in the window; based on the first convolution calculation result, determines the first obstacle recognition result of the overhead rail robot, so as to efficiently identify whether there are obstacles in the target area of ​​the overhead rail robot's operation, further reduce the risk of safety accidents, and improve the safety of the production process.

[0077] Based on any of the above embodiments, the formula of the environment model is as follows: in, , is the origin of the radar device in world coordinates obtained from the controller. To adjust the sign of the direction, is the radar laser ranging value, is the horizontal angle of the radar ray, is the vertical angle of the radar, is the vertical rotation angle of the motor, The coordinates of the data points in world coordinates obtained by aligning the converted radar scan.

[0078] Based on any of the above embodiments, the obstacle recognition result determination unit 530 is specifically configured to: Determining a result score based on the first convolution calculation result, and determining that the first obstacle recognition result is that an obstacle exists when the result score is greater than a threshold score; The threshold score is obtained through training based on first sample data with obstacles, second sample data without obstacles, and label results.

[0079] Based on any of the above embodiments, the target area data determination unit 520 includes a moving unit, and the moving unit is specifically used to: The working process of the Skyrail robot includes a moving process; A warning area determination unit is used to determine the target area data based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot when the overhead rail robot is in the moving process, and the target area data is the warning area data; the first obstacle coordinates and the second obstacle coordinates are the obstacle world coordinate system coordinates obtained by two consecutive scans respectively.

[0080] Based on any of the above embodiments, it further includes an adjustment unit, wherein the adjustment unit is specifically used to: When the second obstacle recognition result corresponding to the target area data indicates that an obstacle exists, controlling the overhead rail robot to decelerate until it stops, and issuing a warning signal; The relative position of the obstacle and the avoidance component of the overhead rail robot is determined, and the avoidance component is adjusted based on the relative position.

[0081] Based on any of the above embodiments, it further includes an in-and-out pool unit, wherein the in-and-out pool unit is specifically used for: Using the cellar area as the alarm area; Performing a sliding window traversal on the alarm area features corresponding to the alarm area to obtain a third convolution calculation result of the data in the window; Based on the third convolution calculation result, a third obstacle recognition result of the overhead rail robot is determined, and when the third obstacle recognition result indicates that an obstacle exists, an alarm signal is issued.

[0082] Based on any of the above embodiments, the target area data determination unit 520 is specifically used to: Determining first target area data based on the environment model and the obstacle world coordinate system coordinates; Filling the missing data or abnormal data in the first target area data to obtain second target area data; The second target area data is sampled by using one of group sampling and cluster sampling to obtain the target area data.

[0083] Figure 6 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 6As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the obstacle recognition method of the overhead rail robot, the method comprising: performing obstacle recognition based on the laser radar point cloud data, determining the environment model in the world coordinate system, and the world coordinate system coordinates of the obstacles around the overhead rail robot; determining the target area data based on the environment model and the world coordinate system coordinates of the obstacles, performing a sliding window traversal on the target area features corresponding to the target area data, and obtaining the first convolution calculation result of the data in the window; determining the first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0084] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0085] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the overhead rail robot obstacle recognition method provided by the above-mentioned methods, and the method includes: performing obstacle recognition based on lidar point cloud data, determining an environment model in a world coordinate system, and the world coordinate system coordinates of obstacles around the overhead rail robot; determining target area data based on the environment model and the world coordinate system coordinates of the obstacle, performing a sliding window traversal on the target area features corresponding to the target area data, and obtaining a first convolution calculation result of the data in the window; and determining a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0086] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the overhead rail robot obstacle recognition method provided by the above-mentioned methods, the method comprising: performing obstacle recognition based on lidar point cloud data, determining an environment model in a world coordinate system, and the world coordinate system coordinates of obstacles around the overhead rail robot; determining target area data based on the environment model and the world coordinate system coordinates of the obstacle, performing a sliding window traversal on the target area features corresponding to the target area data, and obtaining a first convolution calculation result of the data in the window; and determining a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

[0087] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0088] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying obstacles of an overhead rail robot, characterized in that: include: Obstacle identification is performed based on the LiDAR point cloud data to determine the environment model in the world coordinate system and the world coordinate system coordinates of the obstacles around the Tiangan robot; Based on the environment model and the coordinates of the obstacle world coordinate system, determine the target area data, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window; Based on the first convolution calculation result, a first obstacle recognition result of the overhead rail robot is determined.

2. The obstacle recognition method for an overhead rail robot according to claim 1, characterized in that: The formula of the environmental model is as follows: ； ； ； in, , is the origin of the radar device in world coordinates obtained from the controller. To adjust the sign of the direction, is the radar laser ranging value, is the horizontal angle of the radar ray, is the vertical angle of the radar, is the vertical rotation angle of the motor, The coordinates of the data points in world coordinates obtained by aligning the converted radar scan.

3. The obstacle recognition method for an overhead rail robot according to claim 1, characterized in that: The determining, based on the first convolution calculation result, a first obstacle recognition result of the overhead rail robot includes: Determining a result score based on the first convolution calculation result, and determining that the first obstacle recognition result is that an obstacle exists when the result score is greater than a threshold score; The threshold score is obtained through training based on first sample data with obstacles, second sample data without obstacles, and label results.

4. The obstacle recognition method for an overhead rail robot according to any one of claims 1 to 3, characterized in that: The determining of target area data based on the environment model and the obstacle world coordinate system coordinates includes: The working process of the Skyrail robot includes a moving process; When the overhead rail robot is in the moving process, the target area data is determined based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot, and the target area data is the alarm area data; the first obstacle coordinates and the second obstacle coordinates are the obstacle world coordinate system coordinates obtained by two consecutive scans, respectively.

5. The obstacle recognition method for an overhead rail robot according to claim 4, characterized in that: The target area data is determined based on the environment model, the first obstacle coordinates and the second obstacle coordinates, and the moving speed and moving direction of the overhead rail robot, and then further includes: When the second obstacle recognition result corresponding to the target area data indicates that an obstacle exists, controlling the overhead rail robot to decelerate until it stops, and issuing a warning signal; The relative position of the obstacle and the avoidance component of the overhead rail robot is determined, and the avoidance component is adjusted based on the relative position.

6. The obstacle recognition method for an overhead rail robot according to claim 4, characterized in that: The working process of the Skyrail robot also includes the process of entering and exiting the pool; Using the cellar area data as the alarm area data; Performing a sliding window traversal on the alarm area features corresponding to the alarm area data to obtain a third convolution calculation result of the data in the window; Based on the third convolution calculation result, a third obstacle recognition result of the overhead rail robot is determined, and when the third obstacle recognition result indicates that an obstacle exists, an alarm signal is issued.

7. The obstacle recognition method for an overhead rail robot according to any one of claims 1 to 3, characterized in that: The determining of target area data based on the environment model and the obstacle world coordinate system coordinates includes: Determining first target area data based on the environment model and the obstacle world coordinate system coordinates; Filling the missing data or abnormal data in the first target area data to obtain second target area data; The second target area data is sampled by using one of group sampling and cluster sampling to obtain the target area data.

8. An obstacle recognition device for an overhead rail robot, characterized in that: include: A determination unit, used to identify obstacles based on the laser radar point cloud data, determine the environment model in the world coordinate system, and the world coordinate system coordinates of obstacles around the Tiangan robot; A target area data determination unit is used to determine the target area data based on the environment model and the coordinates of the obstacle world coordinate system, perform a sliding window traversal on the target area features corresponding to the target area data, and obtain a first convolution calculation result of the data in the window; The obstacle recognition result determination unit is used to determine a first obstacle recognition result of the overhead rail robot based on the first convolution calculation result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the obstacle recognition method for the overhead rail robot as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the obstacle recognition method for an overhead rail robot as claimed in any one of claims 1 to 7 is implemented.

Images