Moving object clamping stagnation identification method, device, equipment and medium

Abstract

The invention discloses a moving object clamping stagnation identification method, device and equipment and a medium, and belongs to the field of robots. The method comprises the following steps: acquiring a current position of a moving object in a current frame image relative to an SLAM map and a previous position of a moving object in a previous frame image relative to the SLAM map; determining the current moving speed of the moving object according to the current position and the previous position; within a preset time period, determining a t-v fitting result of the moving object according to the current moving speed of the moving object; and judging whether the moving object is blocked or not according to the t-v fitting result in the preset time period. The device comprises a position obtaining module. A moving speed obtaining module; a fitting result obtaining module; and a judgment result obtaining module. According to the technical scheme, the problem of parcel clamping stagnation identification in the production line patrol process can be solved, parcel clamping stagnation processing is facilitated, the system intelligence level is improved, and the manual workload is reduced.

Description

Technical Field

[0001] The present invention relates to the field of robotics, and in particular to a method, device, equipment and medium for identifying a stuck moving object. Background Art

[0002] In the express logistics industry, some companies use advanced intelligent sorting systems. In this system, parcels are automatically sorted after passing through the conveyor belt, and the sorted parcels are then transported to various target areas through the conveyor belt, where they are packed and then stored and transported. In this workflow, a large number of conveyor belts are used to transport parcels. In the actual work process, parcels may get stuck at the joints and turns of the conveyor belts. When parcels get stuck, they need to be handled, otherwise the delivery of the parcels will be affected.

[0003] There are currently two main methods for handling package jams. One method is to use cameras for monitoring, and after manually identifying the jam, ask humans to handle it on the spot. This method is technically easy to implement, but requires real-time manual monitoring, which wastes precious human resources.

[0004] Another method is to deal with package jams through visual detection. This method uses a fixed camera to collect and analyze the package transportation situation, and compares the position of the package in the collected image to determine whether the package is stuck. The camera of this solution is fixed, and the scene that can be monitored and identified is fixed. If a jam is detected, it still needs to be handled manually on site. Summary of the invention

[0005] In order to solve the problems existing in the prior art, the present invention provides a method for identifying a stuck moving object applied to a mobile inspection robot, which comprises:

[0006] Obtain the current position of the moving object in the current frame image relative to the SLAM map and the previous position of the moving object in the previous frame image relative to the SLAM map; determine the current moving speed of the moving object according to the current position and the previous position; determine the TV fitting result of the moving object according to the current moving speed of the moving object within a preset time period; and determine whether the moving object is stuck according to the TV fitting result within the preset time period.

[0007] On the other hand, the present invention provides a mobile object stuck identification device, which includes: a position obtaining module, used to obtain the current position of the mobile object in the current frame image relative to the SLAM map and the previous position of the mobile object in the previous frame image relative to the SLAM map; a moving speed obtaining module, used to determine the current moving speed of the mobile object according to the current position and the previous position; a fitting result obtaining module, used to determine the TV fitting result of the mobile object according to the current moving speed of the mobile object within a preset time period; a judgment result obtaining module, used to judge whether the mobile object is stuck according to the TV fitting result within the preset time period.

[0008] In yet another aspect of the present invention, an electronic device is provided, comprising: a processor and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the above-mentioned method for identifying a stuck moving object.

[0009] On the other hand, the present invention provides a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned moving object stuck identification method.

[0010] The technical solution provided by the embodiment of the present invention has the following beneficial effects:

[0011] By obtaining the current position of the moving object in the current frame image relative to the SLAM map and the previous position of the moving object in the previous frame image relative to the SLAM map; determining the current moving speed of the moving object according to the current position and the previous position; determining the TV fitting result of the moving object according to the current moving speed of the moving object within a preset time period; judging whether the moving object is stuck according to the TV fitting result within the preset time period, the mobile inspection robot can identify the situation of package jams, so that the mobile inspection robot can handle the stuck packages, which is beneficial to improving the intelligence level of the intelligent sorting system and reducing manual workload. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a flow chart of a method for identifying a stuck moving object provided by Embodiment 1 of the present invention;

[0013] Figure 2 is a flow chart of another method for identifying a stuck moving object provided in Embodiment 1 of the present invention;

[0014] Figure 3 This is a posture transformation schematic diagram provided by Embodiment 1 of the present invention;

[0015] Figure 4is a schematic diagram of a package being stuck provided by the first embodiment of the present invention;

[0016] Figure 5 is a schematic diagram of a fitting result provided in Embodiment 1 of the present invention;

[0017] Figure 6 It is a schematic diagram of the structure of a moving object stuck identification device provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0018] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0019] See also Figure 1 , an embodiment of the present invention provides a method for identifying a stuck mobile object, which is suitable for the inspection of a mobile inspection robot, that is, the method is applied to a mobile inspection robot, and when the mobile inspection robot uses the method, it is in a moving state. When the mobile object is a package on a conveying production line (such as a conveyor belt), the mobile inspection robot equipped with the method will patrol the conveying production line for conveying the package to identify whether the package in transit is stuck. The mobile inspection robot has SLAM (Simultaneous Localization and Mapping) technology, and a SLAM map is constructed by this technology. The construction method can be based on vision, radar, or a combination of vision and radar. This embodiment does not specifically limit the technology for constructing a SLAM map. The mobile inspection robot is equipped with a laser radar, through which the laser radar data of the mobile inspection robot's surrounding environment can be obtained; the mobile inspection robot is also equipped with a 3D camera, through which the image data and camera point cloud data of the mobile inspection robot's surrounding environment can be obtained, and these data can be called 3D camera data; the mobile inspection robot is also equipped with an IMU (Inertial Mersurement Unit), which includes: a three-axis accelerometer and a three-axis gyroscope, which can measure the acceleration and angular velocity information of the mobile inspection robot respectively, and these information can be called IMU data. Fusion of these data, using a combination of laser inertial navigation and visual inertial navigation to construct a SLAM map. The process of this method includes the following steps:

[0020] Step S101: obtaining the current position of the moving object in the current frame image relative to the SLAM map and the previous position of the moving object in the previous frame image relative to the SLAM map.

[0021] Obtain the current position of the moving object in the current frame image relative to the SLAM map, that is, obtain the current position of the moving object in the current frame image in the SLAM map. The implementation method of this step can be: the laser radar on the mobile inspection robot will locate its position and posture in the SLAM map in real time, and the 3D camera will collect and analyze in real time whether there are moving objects (such as packages) in the scene around the mobile inspection robot. If there are moving objects, the posture of the center of the moving object relative to the camera coordinate system is calculated. Since the position of the laser radar and the position of the 3D camera are relatively fixed, the relative position relationship can be calibrated in advance. Based on the posture relationship of the moving object in the camera coordinate system, with the help of the posture relationship between the laser radar and the 3D camera and the posture of the laser radar relative to the SLAM map, it is converted into the posture of the moving object relative to the SLAM map, that is, the current position of the moving object in the current frame image in the SLAM map is obtained. It should be noted that the posture of the moving object relative to the SLAM map includes the current position of the moving object in the current frame image in the SLAM map.

[0022] In some embodiments, the following steps (1)-(3) may be included.

[0023] Step (1), according to the relationship between the camera coordinate system and the laser radar coordinate system, determine the relationship between the position and posture of the moving object in the SLAM map in the camera coordinate system.

[0024] Since the global point cloud map, laser radar, and 3D camera are in different coordinate systems, and the mobile inspection robot is in a moving state, it is necessary to convert the position and posture of the moving object in the camera coordinate system. In other words, the coordinate system obtained by SLAM mapping is defined as the world coordinate system. Since the 3D camera obtains the position and posture of the moving object in the camera coordinate system, in order to convert the position and posture of the moving object in the camera coordinate system to the position and posture of the moving object in the world coordinate system, a series of coordinate transformations are required.

[0025] Specifically, Figure 3 As shown in the figure, the position and posture of the moving object in the SLAM map in the camera coordinate system is obtained by the following relationship:

[0026]

[0027] in, It is the pose transformation matrix of the moving object in the SLAM map; is the pose transformation matrix of the laser radar coordinate system relative to the SLAM map coordinate system; is the pose transformation matrix of the camera coordinate system relative to the lidar coordinate system; is the pose transformation matrix of the moving object in the camera coordinate system.

[0028] It should be noted that: since the pose contains position information and attitude information, when the pose transformation matrix is ​​obtained Afterwards, the position can be obtained from it. It is a 4×4 matrix. The first three rows of the fourth column of the matrix are the positions of the moving object in the camera coordinate system in the world coordinate system, that is, the position in the SLAM map; Obtained by pre-calibration of LiDAR and 3D camera.

[0029] Step (2), obtaining the position and posture of the moving object in the current frame image in the camera coordinate system.

[0030] First, by performing target recognition and segmentation on the two-dimensional image data output by the 3D camera, segmentation data (or target recognition results) are obtained, and the segmentation data is used to characterize the moving object. In practical applications, the YOLACT (YouOnly Look At CoefficienTs, real-time instance segmentation model) model can be used to process the two-dimensional image data of the 3D camera to obtain the target recognition result. The target recognition result is in the form of a matrix mask, and each element in the matrix represents whether the pixel corresponding to the element belongs to a moving object. If the value of an element in the mask matrix is ​​1, it means that the pixel corresponding to the element is part of the target to be measured (i.e., the moving object to be measured); if it is 0, it means that it has nothing to do with the target to be measured (i.e., the moving object to be measured).

[0031] Then, the camera point cloud data is segmented according to the target recognition result to obtain the target point cloud data. Since the 3D camera can obtain image data and camera point cloud data aligned with the coordinate system, the camera point cloud data is segmented according to the target recognition result to obtain the initial target point cloud data containing the moving object. After that, the point cloud data is subjected to denoising and downsampling, and the obtained point cloud data is the target point cloud data.

[0032] Finally, the target point cloud data is processed by the minimum bounding box method to obtain the minimum bounding box of the moving object, and the center of the minimum bounding box is calculated, which is the coordinate value of the geometric center of the moving object. Figure 4 As shown, the present invention processes the target point cloud data through the OBB (Oriented Bounding Box) method to obtain the target bounding box, and then calculates the center coordinate value of the target bounding box, wherein the OBB method is a prior art and the present invention does not make a detailed description of it. In the OBB method, the three mutually perpendicular edges intersecting at the same vertex respectively construct the object posture direction vector in order of length. The three posture direction vectors of the moving object and the center coordinates of the moving object constitute the posture transformation matrix of the moving object in the camera coordinate system. The matrix is ​​a 4×4 matrix. The first three columns and three rows of the matrix are the posture of the moving object in the camera coordinate system, and the first three rows of the fourth column represent the position coordinates of the object in the camera coordinate system.

[0033] The pose transformation matrix obtained through the above steps is the pose of the moving object in the camera coordinate system.

[0034] Step (3), obtaining the current posture according to the relationship and the posture of the moving object in the camera coordinate system in the current frame image.

[0035] According to the relationship obtained in step (1) and the pose of the moving object in the camera coordinate system obtained in step (2), the current pose can be obtained, that is, the current pose of the moving object in the current frame image relative to the SLAM map is obtained, and correspondingly, the previous pose of the moving object in the previous frame image relative to the SLAM map is obtained. The method for obtaining the previous pose can refer to the method for obtaining the current pose, and this embodiment will not be repeated.

[0036] It should be noted that: based on the relationship obtained in step (1), it is also necessary to obtain the position and posture of the laser radar (i.e., the mobile inspection robot) in the SLAM map. The method for obtaining the position and posture can be as follows:

[0037] First, it is necessary to obtain a global point cloud map of the environment where the mobile inspection robot is located according to the laser radar data, 3D camera data and IMU data, and then estimate the position and posture of the mobile inspection robot. The present invention constructs a global point cloud map by the method of LVI-SAM (lindar-vision-imu slam, laser visual inertial navigation SLAM system). LVI-SAM is based on factor graphs and consists of two subsystems: a visual-inertial system and a laser radar-inertial system. The mapping method based on laser radar can obtain detailed information of a long-distance environment, but is not suitable for unstructured environments, such as corridors or flat open areas; the mapping method based on vision is suitable for scene recognition, but the change of illumination and rapid movement have a great influence on the effect of mapping. Based on this, the LVI-SAM method improves the robustness and accuracy of the entire system by constructing a tightly coupled visual-inertial system and a laser radar-inertial system.

[0038] Then, the LiDAR data and the global point cloud map are registered by the point cloud registration method to obtain the position and posture of the mobile inspection robot in the SLAM map. The position and posture of the mobile inspection robot in the scene is constantly changing, and the data obtained by the LiDAR in real time is the scene information within a limited range. Therefore, it is necessary to register the LiDAR data with the global point cloud map to obtain the current position and posture of the mobile inspection robot in the scene, that is, the position and posture of the mobile inspection robot.

[0039] In some possible implementations, it is necessary to obtain the initial estimated pose of the mobile inspection robot. The initial estimated pose can be calculated based on the IMU data or set by the user. After that, the current frame lidar data and the global point cloud map are aligned to obtain the pose of the mobile inspection robot in the SLAM map. In practical applications, the point cloud alignment can be performed using the NDT (Normal Distributions Transform) method. First, the global point cloud map is rasterized and the normal distribution parameters are calculated. Then, the current frame lidar data is aligned and the global optimal solution is obtained based on the initial estimated pose. After the above steps, the pose of the mobile inspection robot, that is, the pose of the moving object in the radar coordinate system, can be obtained.

[0040] This embodiment does not limit the specific implementation method of this step.

[0041] Step S102: Determine the current moving speed of the moving object according to the current position and the previous position.

[0042] First, according to step S101, and in For i The current position of the moving object in the current frame image corresponding to the moment relative to the SLAM map; For i-1 The last position of the moving object relative to the SLAM map in the last frame image corresponding to the moment.

[0043] get and Then, the current moving speed of the moving object is obtained according to the following formula:

[0044]

[0045] in, is the current moving speed of the moving object; (t i -t i-1 ) is the time interval between two adjacent frames of data collection. The collected data includes lidar data and 3D camera data, where the 3D camera data includes two-dimensional image data and three-dimensional point cloud data. That is, at t i Collect the laser radar data and 3D camera data at the same time, and then obtain the corresponding In some embodiments, the time interval is 50 ms.

[0046] Step S103: within a preset time period, determining the tv fitting result of the moving object according to the current moving speed of the moving object.

[0047] Since the obtained geometric center coordinates of the moving object have deviations, the calculated current moving speed of the moving object will fluctuate. Therefore, the present invention analyzes the data by using the Ransac (Random Sample And Consensus) method, that is, the current moving speeds (t i ,v i ) is used for statistics. It should be noted that v i for The preset number can be 1000. After downsampling, the moving speed data set is obtained. Then, a two-dimensional coordinate system is constructed with t as the horizontal coordinate and v as the vertical coordinate. The moving speed data set is fitted by the Ransac method.

[0048] Specifically, Figure 5 As shown, we randomly select two points (t i ,v i )、(t j ,v j ) calculates a straight line passing through two points that satisfies the equation:

[0049] Ax+By+C=0 (3)

[0050] Among them, A, B, and C are the coefficients of the straight line equation obtained from the two points, and then the points other than the above two points in the moving speed data set are calculated (t q ,v q ) is the sum of the distances d from the above straight line sum , the distance calculation formula is as follows:

[0051]

[0052] Among them, d sum is the distance d q The sum of q For point (t q ,v q ) from the straight line.

[0053] Repeat the above steps to obtain a set of straight lines and their corresponding d sum , and finally select the sum of the distances d sum The smallest straight line is the final TV fitting result.

[0054] Step S104: determining whether the moving object is stuck according to the tv fitting result within a preset time period.

[0055] After obtaining the fitting result of step S104, it is possible to judge whether the target object (ie, the moving object) has entered a stuck state. The specific judgment process is: setting judgment conditions, including a first judgment condition, a second judgment condition, and a third judgment condition.

[0056] The first judgment condition is: the slope k of the straight line of the fitting result is within the first preset threshold range, where the first preset threshold range can be |k|<0.2, or other reasonable threshold ranges. The second judgment condition is: the intersection h of the straight line of the fitting result and the y-axis is greater than the second preset threshold, and can also be greater than the second preset threshold range and less than the fourth preset threshold, where the second preset threshold can be 0.8m / s, the fourth preset threshold can be 1.5m / s, and can also be other reasonable numerical ranges set according to actual conditions. The third judgment condition is: the intersection h of the straight line of the fitting result and the y-axis is within the third preset threshold range, which can be |h|<0.1.

[0057] Then, the fitting result is judged according to the first judgment condition, the second judgment condition and the third judgment condition. The specific judgment process is as follows:

[0058] First, the fitting result is judged by the first judgment condition. If it passes, it indicates that the moving object is in a uniform motion state or a stationary state, and then the subsequent judgment process continues; if it fails, it indicates that the moving object is in an acceleration or deceleration state, and the fitting result cannot be used as a subsequent judgment process and is considered invalid;

[0059] The fitting result is then judged using the second judgment condition. If the intersection h of the straight line of the fitting result and the y-axis is within the second preset threshold range, a judgment result is obtained that the moving object has not entered a stuck state; if the intersection h of the straight line of the fitting result and the y-axis is within the third preset threshold range, a judgment result is obtained that the moving object has entered a stuck state.

[0060] Through the above judgment results, a judgment result of whether the moving object enters a stuck state is obtained.

[0061] During the inspection process of the mobile inspection robot, the method for identifying the stuck moving object also includes:

[0062] In response to the mobile object being stuck, a processing instruction is generated, where the processing instruction is used to instruct the mobile inspection robot to process the mobile object.

[0063] In this way, when the mobile inspection robot is inspecting the production line, if it finds that a moving object is stuck, it can be driven to process the stuck moving object, further realizing unmanned sorting, improving the level of intelligence, and reducing manual workload.

[0064] See also Figure 6An embodiment of the present invention provides a moving object stuck identification device, which is applied to a mobile inspection robot. The device includes: a position obtaining module 201, a moving speed obtaining module 202, a fitting result obtaining module 203, and a judgment result obtaining module 204.

[0065] Among them, the position obtaining module 201 is used to obtain the current position of the moving object in the current frame image relative to the SLAM map and the previous position of the moving object in the previous frame image relative to the SLAM map; the moving speed obtaining module 202 is used to determine the current moving speed of the moving object according to the current position and the previous position; the fitting result obtaining module 203 is used to determine the tv fitting result of the moving object according to the current moving speed of the moving object within a preset time period; the judgment result obtaining module 204 judges whether the moving object is stuck according to the fitting result within the preset time period.

[0066] Optionally, the position acquisition module 201 is specifically used to: determine the relationship between the posture of the moving object in the camera coordinate system in the SLAM map according to the camera coordinate system and the lidar coordinate system; obtain the posture of the moving object in the camera coordinate system in the current frame image; and obtain the current posture according to the relationship and the posture of the moving object in the camera coordinate system in the current frame image, wherein the current posture includes the current position.

[0067] Optionally, the position obtaining module 201 includes a current position obtaining module, in which the position and posture of the moving object in the current frame image in the camera coordinate system are obtained, specifically: target recognition and segmentation are performed on the moving object in the current frame image to obtain segmentation data; segmentation is performed according to the segmentation data and the three-dimensional point cloud data containing the moving object to obtain the target point cloud data of the moving object; the target point cloud data is processed using a minimum bounding box algorithm to obtain the center coordinates and posture of the minimum bounding box, the center coordinates are used as the center coordinates of the moving object, and the bounding box posture is used as the posture of the moving object, to construct the position and posture matrix of the moving object in the camera coordinate system; and the posture matrix is ​​used as the position and posture of the moving object in the camera coordinate system.

[0068] Optionally, the moving speed obtaining module 202 is specifically used to determine the current moving speed of the moving object according to the following formula: in, is the current moving speed of the moving object; For i The current position of the moving object in the current frame image corresponding to the moment relative to the SLAM map; For i-1 The previous position of the moving object relative to the SLAM map in the previous frame image corresponding to the moment.

[0069] Optionally, the fitting result obtaining module 203 is specifically used to: construct a data set within a preset time period, the data set having: a plurality of data points, the data points including the current moving speed and the time corresponding to the current moving speed; arbitrarily select two data points (t i ,v i )、(t j ,v j ), calculate the equation of the line passing through the two data points, the formula is as follows: Ax+By+C=0, where A, B, and C are the coefficients of the line equation respectively; v i and v j The moving speed and scalar form; calculate the data points (t q ,v q ) The sum of the distances from the straight line d sum , the distance calculation formula is as follows: Among them, d q is the data point (t q ,v q ) from the straight line; traverse all any two data points to obtain the set of straight line equations and the d corresponding to each straight line equation in the set of straight line equations sum , select the smallest d from the set of line equations sum The corresponding straight line is the TV fitting result within the preset time period.

[0070] Optionally, the judgment result obtaining module 204 is specifically used to: judge whether the straight line slope of the TV fitting result meets the first judgment condition, if so, judge that the TV fitting result is valid, otherwise, judge that the TV fitting result is invalid; when the TV fitting result is valid, judge whether the intercept of the TV fitting result meets the second judgment condition, if so, judge that the moving object is not stuck, otherwise, judge whether the intercept meets the third judgment condition, if so, judge that the moving object is stuck; wherein the first judgment condition is within the first preset threshold range, the second judgment condition is greater than the second preset threshold, and the third judgment condition is within the third preset threshold range.

[0071] Optionally, the second judgment condition is greater than a second preset threshold and less than a fourth preset threshold.

[0072] Optionally, the device further comprises a processing instruction generating module, which is used to generate a processing instruction in response to a mobile object getting stuck during the inspection process of the mobile inspection robot, wherein the processing instruction is used to instruct the mobile inspection robot to process the mobile object.

[0073] It should be noted that: the mobile object jam identification device provided in the above embodiment only uses the division of the above functional modules as an example to illustrate when identifying the jam of a mobile object. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the mobile object jam identification device provided in the above embodiment and the mobile object jam identification method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0074] An embodiment of the present invention provides an electronic device, which includes: a memory and a processor. The processor is connected to the memory and is configured to execute the above-mentioned moving object stuck identification method based on the instructions stored in the memory. The number of processors can be one or more, and the processor can be single-core or multi-core. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip. The memory can be an example of the following computer-readable medium.

[0075] The embodiment of the present invention provides a computer-readable storage medium, on which at least one instruction, at least one program, code set or instruction set is stored, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned mobile object jam identification method. Computer-readable storage media include: permanent and non-permanent, removable and non-removable media can implement information storage by any method or technology. Information can be a computer-readable instruction, a data structure, a module of a program or other data. Examples of computer storage media include, but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk-read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassette, disk storage or other magnetic storage device or any other non-transmission medium, which can be used to store information that can be accessed by a computing device.

[0076] It is known from common technical knowledge that the present invention can be implemented by other embodiments that do not deviate from its spirit or essential features. Therefore, the above disclosed embodiments are only illustrative in all respects and are not exclusive. All changes within the scope of the present invention or within the scope equivalent to the present invention are included in the present invention.

Claims

1. A method for identifying a stuck moving object, characterized in that: The mobile object stuck identification method is applied to a mobile inspection robot, the mobile inspection robot is constructed with a SLAM map, and the method comprises: Obtaining a current position of a moving object in a current frame image relative to the SLAM map and a previous position of the moving object in a previous frame image relative to the SLAM map; Determining a current moving speed of the moving object according to the current position and the previous position; Determining a TV fitting result of the moving object according to the current moving speed of the moving object within a preset time period; Whether the moving object is stuck is determined according to the tv fitting result within the preset time period.

2. A method for identifying a stuck moving object according to claim 1, characterized in that: The obtaining of the current position of the moving object in the current frame image relative to the SLAM map comprises: According to the relationship between the camera coordinate system and the laser radar coordinate system, a relationship equation of the position and posture of the moving object in the camera coordinate system on the SLAM map is determined; The current posture is obtained according to the relationship and the posture of the moving object in the current frame image in the camera coordinate system, and the posture includes the position.

3. A method for identifying a stuck moving object according to claim 2, characterized in that: The step of obtaining the position and posture of the moving object in the current frame image in the camera coordinate system includes: Performing target recognition and segmentation on the moving object in the current frame image to obtain segmentation data; Segmenting the moving object based on the segmentation data and the camera three-dimensional point cloud data containing the moving object to obtain target point cloud data of the moving object; Processing the target point cloud data using a minimum bounding box algorithm to obtain a minimum bounding box posture, wherein the center coordinates included in the posture are used as the center coordinates of the moving object; The center coordinates of the moving object are used as the position of the moving object in the camera coordinate system.

4. A method for identifying a stuck moving object according to claim 1, characterized in that: Determining the current moving speed of the moving object according to the current position and the previous position includes: The current moving speed of the moving object is determined according to the following formula: in, is the current moving speed of the moving object; For i The current position of the moving object in the current frame image corresponding to the moment relative to the SLAM map; For i-1 The last position of the moving object relative to the SLAM map in the last frame image corresponding to the moment.

5. A method for identifying a stuck moving object according to claim 1, characterized in that: Determining the TV fitting result of the moving object according to the current moving speed of the moving object within the preset time period includes: Within a preset time period, constructing a data set, the data set having: a plurality of data points, the data points including a current moving speed and a time corresponding to the current moving speed; In the data set, any two data points (t i ,v i )、(t j ,v j ), calculate the equation of the straight line passing through the two data points, the formula is as follows: Ax+By+C=0, Where A, B, and C are the coefficients of the linear equation respectively; v i and v j The moving speeds are and The scalar form of ; Calculate the data points (t q ,v q ) is the sum of the distances d from the straight line sum , the distance calculation formula is as follows: Among them, d q is the data point (t q ,v q ) from the said straight line; Traverse all arbitrary two data points to obtain the set of straight line equations and the d corresponding to each straight line equation in the set of straight line equations sum , select the smallest d from the set of straight line equations sum The corresponding straight line is the tv fitting result within the preset time period.

6. A method for identifying a stuck moving object according to claim 1, characterized in that: The determining whether the moving object is stuck according to the tv fitting result within the preset time period includes: Determine whether the straight line slope of the TV fitting result meets the first judgment condition, if so, determine that the TV fitting result is valid, otherwise, determine that the TV fitting result is invalid; When the TV fitting result is valid, judging whether the intercept of the TV fitting result satisfies a second judgment condition, if so, judging that the moving object is not stuck, otherwise, judging whether the intercept satisfies a third judgment condition, if so, judging that the moving object is stuck; Among them, the first judgment condition is within a first preset threshold range, the second judgment condition is greater than a second preset threshold, and the third judgment condition is within a third preset threshold range.

7. A method for identifying a stuck moving object according to claim 6, characterized in that: The second judgment condition is greater than the second preset threshold and less than the fourth preset threshold; During the inspection process of the mobile inspection robot, the mobile object stuck identification method further includes: In response to the mobile object being stuck, a processing instruction is generated, where the processing instruction is used to instruct the mobile inspection robot to process the mobile object.

8. A moving object stuck identification device, characterized in that: The device comprises: A position obtaining module, used to obtain the current position of the moving object in the current frame image relative to the SLAM map and the previous position of the moving object in the previous frame image relative to the SLAM map; A moving speed obtaining module, used to determine the current moving speed of the moving object according to the current position and the previous position; A fitting result obtaining module, used to determine the TV fitting result of the moving object according to the current moving speed of the moving object within a preset time period; The judgment result obtaining module judges whether the moving object is stuck according to the tv fitting result within the preset time period.

9. An electronic device, characterized in that: The electronic device comprises: a processor and a memory for storing executable instructions of the processor; The processor is configured to execute the moving object stuck identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the mobile object jam identification method described in any one of claims 1-7.

Images