Intelligent inspection robot for belt conveyor and control method of intelligent inspection robot

Abstract

The invention provides a belt conveyor intelligent inspection robot and a control method thereof, and relates to the field of mechanical automation. The intelligent inspection robot for the belt conveyor comprises a robot body and a position sensor, the position sensor is arranged on the robot body, and the robot body carries a central processor developed based on a robot operating system of an ROS2 framework. The position information of the corresponding robot is obtained after the position information is processed by the central processor, and meanwhile, mapping is performed to obtain corresponding real-time map information, so that autonomous navigation of the robot can be performed after the position information of the robot and the corresponding real-time map information are obtained, obstacles are avoided, and the navigation efficiency of the robot is improved. And meanwhile, the current position of the robot can be obtained in real time, and the corresponding robot is set to move to the corresponding position of the belt conveyor for inspection, so that the practicability is higher than that in the prior art.

Description

Technical Field

[0001] This application relates to the field of mechanical automation. Specifically, it relates to an intelligent inspection robot for a belt conveyor and its control method. Background Art

[0002] In recent years, with the development of industrial technology, the corresponding automated intelligent technology has also developed rapidly. In industrial assembly lines, for complex and repetitive labor, such as monitoring and inspection, etc., robot technology in automated technology can generally be used to solve the problem. Especially in industrial assembly lines using belt conveyors, due to the long length of the belt conveyor equipment and the large range that needs to be inspected and monitored, the labor cost required for manual monitoring is relatively high. Therefore, robots are generally used for inspection.

[0003] In the existing technology, there are two solutions for monitoring belt conveyors in industrial assembly lines. One is to use multiple monitoring robots at fixed positions to monitor the corresponding belt conveyors. Since a large number of monitoring robots are involved, the cost is relatively high. The other is to use robots moving along a set fixed route for inspection and monitoring. Due to the complex industrial site scenarios, there may be existing obstacles in the fixed route. And the robots in the existing technology cannot perform autonomous navigation to avoid obstacles during inspection because the patrol route is fixed. Therefore, the practicality is relatively low.

[0004] In view of the above technology, finding an intelligent inspection robot control method for a belt conveyor with higher practicality is an urgent problem for those skilled in the art. Summary of the Invention

[0005] The purpose of this application is to provide an intelligent inspection robot for a belt conveyor and its control method, which can solve the problem that the robot in the existing technology cannot perform autonomous navigation to avoid obstacles during inspection because the patrol route is fixed.

[0006] This application is implemented as follows:

[0007] In a first aspect, this application provides an intelligent inspection robot for a belt conveyor, including: a robot body and a position sensor. The position sensor is arranged on the robot body, and the robot body is equipped with a central processor developed based on the robot operating system of the ROS2 framework;

[0008] The robot body moves based on the control instructions of the central processor. The robot body is also used to collect the operation information of the belt conveyor to be detected and send the operation information to the central processor for analysis to obtain an analysis result. The central processor sends the analysis result to the control terminal through a preset communication network;

[0009] The robot operating system uses the Cartographer algorithm to obtain the sensing data of the position sensor, estimates the position information of the robot using a Bayesian filter based on the sensing data, and optimizes the position information using a non-linear optimization algorithm. The robot operating system also uses the ORB-SLAM3 algorithm to generate real-time map information through dense mapping based on the sensing data, and the central processor performs autonomous navigation of the robot based on the position information of the robot and the real-time map information.

[0010] Based on the first aspect, the robot body further includes a monitoring device, a chassis, and a power system. The chassis is disposed at the bottom of the outer shell of the robot body, the monitoring device is disposed on the outer shell, the signal output end of the monitoring device is connected to the signal receiving end of the central processor, the driving end of the power system is connected to the chassis, and the control end of the power system is connected to the central processor;

[0011] The power system is configured to receive a control instruction from the central processor and drive the chassis to move based on the control instruction. The monitoring device is configured to collect the operation information of the belt conveyor to be detected and send the operation information to the central processor for analysis to obtain an analysis result.

[0012] Based on the first aspect, it further includes: a suspension and shock absorption device, an environmental sensor, and a safety protection device;

[0013] The suspension and shock absorption device is disposed on the chassis. The environmental sensor is used to monitor gas concentration, pollutants, environmental temperature, and humidity in the environment. The safety protection device includes a collision sensor and an emergency stop button. The safety protection device is connected to the drive motor and is configured to perform emergency braking when an abnormal situation occurs to the robot.

[0014] Based on the first aspect, the preset communication network is a hierarchical structure network built based on a network communication protocol. The preset communication network includes a core network, an aggregation network, and an access network, and uses optical fiber or gigabit Ethernet to access corresponding network communication devices for communication.

[0015] In a second aspect, the present application further provides a control method for a belt conveyor intelligent inspection robot, which is applied to the above-mentioned belt conveyor intelligent inspection robot, and includes the following steps:

[0016] Use the Cartographer algorithm to obtain the sensing data of the position sensor, and estimate the position information of the robot using a Bayesian filter based on the sensing data;

[0017] Use a non-linear optimization algorithm to optimize the position information;

[0018] Use the ORB-SLAM3 algorithm to perform dense mapping based on the sensing data to generate real-time map information;

[0019] Control the robot body to move to a preset position based on the optimized position information and the real-time map information;

[0020] Control the robot body to collect the operation information of the belt conveyor to be detected, and send the operation information to a preset network structure to the control terminal.

[0021] Based on the second aspect, the position sensor includes a lidar, an odometer, and an IMU. The Cartographer algorithm is used to obtain the sensing data of the position sensor, and the position information of the robot obtained by estimating the sensing data using a Bayesian filter includes:

[0022] Obtain the processed sensing data of the position sensor, and perform scan matching on the sensing data. The sensing data includes the point cloud data of the lidar filtered by a voxel filter, the odometer sensing data screened by a pose extrapolator, and the IMU data calibrated by a gravity alignment device;

[0023] Use a Bayesian filter to filter and screen the sensing data, and discard the sensing data that does not meet the preset requirements;

[0024] Build a position submap based on the screened sensing data and update the voxel network;

[0025] Insert the point cloud data filtered by the voxel filter into the voxel network to obtain the position information of the robot.

[0026] Based on the second aspect, using a non-linear optimization algorithm to optimize the position information includes:

[0027] Use a non-linear optimization algorithm to calculate the in-submap constraints and inter-submap constraints of the position submap to obtain the error of the position information;

[0028] Build an optimization model based on the odometer sensing data, the IMU data, and the pose data of the fixed frame;

[0029] Perform pose optimization on the error of the position information based on the optimization model to obtain the optimized position information of the robot.

[0030] Based on the second aspect, the use of the ORB-SLAM3 algorithm to perform dense mapping based on the sensing data to generate real-time map information includes:

[0031] Obtain the depth image information in the sensing data, preprocess the sensing data, perform pose estimation on the depth image information, and establish a local map;

[0032] Create a key frame for the processed depth image information, and insert the key frame into the thread of the local map;

[0033] Perform three-dimensional bad point removal on the key frame, and create new three-dimensional points for local BA optimization to remove redundant key frames;

[0034] Establish a similarity model based on the processed key frames for loop closure fusion, and globally optimize the local map based on the similarity model to obtain the real-time map information.

[0035] Based on the second aspect, after using the ORB-SLAM3 algorithm to perform dense mapping and high-precision positioning based on the sensing data to obtain the real-time map information, it further includes:

[0036] Extract the map data feature points in the sensing data, and match the feature points with a preset map;

[0037] Based on the matching result, detect whether the intelligent belt conveyor inspection robot has passed through the explored area, and update the real-time map information.

[0038] Compared with the prior art, the present application has at least the following advantages or beneficial effects:

[0039] The present application proposes an intelligent belt conveyor inspection robot, including a robot body and a position sensor. The position sensor is arranged on the robot body. The robot body is equipped with a central processor developed based on the robot operating system of the ROS2 framework. Different from the previous solutions, a position sensor is set on the robot in the present application. After the sensing data is processed by the Cartographer algorithm of the central processor, the corresponding position information of the robot is obtained. At the same time, mapping is performed based on the ORB-SLAM3 algorithm to obtain the corresponding real-time map information. After having the position information of the robot and the corresponding real-time map information, the robot can perform autonomous navigation, avoid obstacles, and at the same time can obtain the current position of the robot in real time and set the corresponding robot to move to the corresponding position of the belt conveyor for inspection. Therefore, the practicality is relatively high compared with the prior art.

[0040] The present application also provides an intelligent belt conveyor inspection robot management method, corresponding to the above intelligent belt conveyor inspection robot, so the beneficial effects are the same as above. Description of the Drawings

[0041] To more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0042] Figure 1 It is a flowchart of a control method for an intelligent inspection robot of a belt conveyor provided by an embodiment of the present application;

[0043] Figure 2 It is a flowchart of a control method for an intelligent inspection robot of a belt conveyor provided by an embodiment of the present application;

[0044] Figure 3 It is a flowchart of a method for obtaining position information provided by an embodiment of the present application;

[0045] Figure 4 It is a flowchart of a method for optimizing position information provided by an embodiment of the present application;

[0046] Figure 5 It is a flowchart of a method for generating a real-time map provided by an embodiment of the present application.

[0047] Icons: 1. Robot body; 2. Position sensor; 3. Central processor; 4. Outer shell; 5. Monitoring device; 6. Chassis; 7. Power system. Detailed implementation manners

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Usually, the components of the embodiments of the present application described and shown in the accompanying drawings here can be arranged and designed in various different configurations.

[0049] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application that is required to be protected, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0050] The following will describe some implementation manners of the present application in detail with reference to the accompanying drawings. Without conflict, the following various embodiments and the various features in the embodiments can be combined with each other.

[0051] Embodiment

[0052] Figure 1 This is a structural diagram of an intelligent inspection robot for a belt conveyor provided by an embodiment of the present application. As Figure 1 shown, it includes: a robot body 1 and a position sensor 2. The position sensor 2 is arranged on the robot body 1, and the robot body 1 is equipped with a central processor developed based on the robot operating system of the ROS2 framework;

[0053] The robot body 1 moves based on the control instructions of the central processor. The robot body 1 is also used to collect the operation information of the belt conveyor to be detected, and send the operation information to the central processor 3 for analysis to obtain an analysis result. The central processor 3 sends the analysis result to the control terminal through a preset communication network;

[0054] To achieve the above effects, in some embodiments, the robot body 1 further includes a housing 4, a monitoring device 5, a chassis 6 and a power system 7. The power system 7 is used to receive the control instructions of the central processor 3, and drive the chassis 6 to drive the housing 4 to move based on the control instructions. The monitoring device 5 is used to collect the operation information of the belt conveyor to be detected, and send the operation information to the central processor 3 for analysis to obtain an analysis result, and send the analysis result to the control terminal through a preset communication network;

[0055] The central processor 3 is equipped with a robot operating system based on the ROS2 framework. The robot operating system uses the Cartographer algorithm to obtain the sensing data of the position sensor 2, uses the Bayesian filter for estimation based on the sensing data to obtain the position information of the robot, and uses a non-linear optimization algorithm to optimize the position information. The robot operating system also uses the ORB-SLAM3 algorithm to generate real-time map information based on the sensing data for dense mapping. The central processor 3 performs autonomous navigation of the robot based on the position information of the robot and the real-time map information.

[0056] It should be noted that in this embodiment, the specific operation structure of the corresponding robot body 1 is not limited, Figure 1 but only a structural solution of a robot provided by an embodiment of the present application, and is not limited by Figure 1 the content shown therein for the robot structure in this embodiment.

[0057] It should be noted that the robot operating system is developed based on the ROS2 (Robot Operating System 2) framework, and Table 2 is a characteristic table of the corresponding ROS framework.

[0058] Table 2: Characteristic table of the ROS framework

[0059]

[0060]

[0061] It can be understood that for the above-mentioned Cartographer algorithm, this algorithm has advantages such as high precision, high efficiency, and scalability. In this embodiment, the specific method of using the Cartographer algorithm to estimate the position information of the corresponding robot is not specifically limited. At the same time, the specific method of using ORB-SLAM3 to draw a real-time map is not specifically limited, and in this embodiment, the preset network structure for communication with the control terminal is not limited.

[0062] This embodiment proposes an intelligent inspection robot for a belt conveyor, including a robot body 1 and a position sensor 2. The position sensor 2 is arranged on the robot body 1. The robot body 1 is equipped with a central processor 3 developed based on the robot operating system of the ROS2 framework. In this embodiment, a position sensor 2 is set on the robot. After the sensing data of the Cartographer algorithm is processed by the central processor 3, the corresponding position information of the robot is obtained. At the same time, a map is built based on the ORB-SLAM3 algorithm to obtain the corresponding real-time map information. After having the position information of the robot and the corresponding real-time map information, the robot can perform autonomous navigation to avoid obstacles. At the same time, the current position of the robot can be obtained in real time and the corresponding robot can be set to move to the corresponding position of the belt conveyor for inspection. Therefore, the practicability is relatively high compared with the prior art.

[0063] In the above embodiment, the specific structure of the robot body 1 is not limited, such as Figure 1 As shown, exemplarily, in some embodiments of the present application, the robot body 1 further includes a monitoring device 5, a chassis 6, and a power system 7. The chassis 6 is arranged at the bottom of the outer shell 4 of the robot body 1. The monitoring device is arranged on the outer shell 4. The signal output end of the monitoring device 5 is connected to the signal receiving end of the central processor 3. The driving end of the power system 7 is connected to the chassis 6, and the control end of the power system 7 is connected to the central processor 3;

[0064] The power system 7 is used to receive the control instruction of the central processor 3 and drive the chassis 6 to drive the outer shell 4 to move based on the control instruction. The monitoring device 5 is used to collect the operation information of the belt conveyor to be detected and send the operation information to the central processor 3 for analysis to obtain an analysis result.

[0065] It should be noted that in this embodiment, the specific types of the outer shell 4, the monitoring device 5, the chassis 6, and the power system 7 of the corresponding robot body 1 are not limited. The following is a proposed robot hardware structure provided by the embodiments of the present application. The chassis 6 is an important part of the robot system, which bears the movement, navigation, and stability performance of the robot.

[0066] The structural design of the robot chassis 6 should consider the size, weight, and load requirements of the robot, as well as the degrees of freedom of motion required by the robot. Based on the existing requirements: orbital inspection, the chassis 6 structure adopts a single-rail or multi-rail design. Table 1 shows the corresponding optional chassis 6 and their respective advantages and disadvantages.

[0067] Table 1: Types of chassis 6 and their advantages and disadvantages

[0068]

[0069]

[0070] As shown in the above table, it can be understood that the single-rail chassis is suitable for scenarios with limited space and low load requirements, while the multi-rail chassis is suitable for scenarios that require strong mobility and high load-bearing capacity. The choice between a single-rail and a multi-rail as the chassis needs to be comprehensively considered according to the specific application requirements and scenario characteristics. Regarding the power system 7, the power system 7 of the robot chassis 6 generally consists of a motor, a battery, and a drive circuit. The motor selects a DC hub motor. The selection of the battery should be determined according to the usage time and load requirements of the robot, and at the same time, the capacity and charging method of the battery need to be considered. In this embodiment, the hardware components for realizing the functions of the above-mentioned robot body 1 are further defined, improving the integrity of the solution.

[0071] Considering that in order to further improve the safety and stability performance of the corresponding robot, exemplarily, in some embodiments of the present application, it further includes: a suspension and shock absorption device, an environmental sensor, and a safety protection device;

[0072] The suspension and shock absorption device is arranged on the chassis 6. The environmental sensor is used to monitor the gas concentration, pollutants, environmental temperature, and humidity in the environment. The safety protection device includes a collision sensor and an emergency stop button. The safety protection device is connected to the drive motor and is used to perform emergency braking when the robot encounters an abnormal situation.

[0073] The suspension and shock absorption system of the robot chassis 6 provided in this embodiment can provide adaptability and stability to the ground, reduce the vibration and bump of the robot during movement, and help improve the recognition efficiency during the inspection process. The safety protection device in this embodiment generally refers to the safety protection measures that need to be considered in the design of the robot chassis 6, such as collision sensors, emergency stop buttons, etc., to ensure that the robot can stop moving in time and protect the people and objects around it when encountering dangerous situations.

[0074] In this embodiment, by designing the corresponding suspension and shock absorption system and safety protection device, the safety and stability of the robot during operation are ensured.

[0075] In the above embodiments, the specific communication network is not limited. Exemplarily, in some embodiments of the present application, the preset communication network is a hierarchical network built based on network communication protocols. The preset communication network includes a core network, an aggregation network, and an access network, and uses optical fiber or gigabit Ethernet to access the corresponding network communication devices for communication.

[0076] It should be noted that, for the network communication protocol, no specific limitation is made in this embodiment. Due to the structural and communication differences of the position sensor 2 and the central processor 3 of the robot, there are also certain differences in the generally adopted network communication protocols. Therefore, the network communication protocol here can include the content of multiple network communication protocols at the same time. Table 3 is a table of commonly used communication protocols on the market and their specific application objects.

[0077] Table 3: Commonly used communication protocols on the market and their specific application objects

[0078]

[0079]

[0080] It can be understood that the corresponding central processor 3 can also train a targeted data model based on cloud computing, big data technology, and deep learning, and can perform real-time analysis and processing on the data of the robot, improving the decision-making ability and intelligent level of the robot; improving the working efficiency and stability of the robot, and improving the accuracy of visual recognition.

[0081] Based on the same inventive concept, Figure 2 is a flowchart of a control method for an intelligent inspection robot of a belt conveyor provided in an embodiment of the present application, which is applied to the intelligent inspection robot of the belt conveyor in the above embodiment, as Figure 2 shown. The method includes the following steps:

[0082] S10: Use the Cartographer algorithm to obtain the sensing data of the position sensor, and use the Bayesian filter to estimate the position information of the robot based on the sensing data;

[0083] S11: Optimize the position information using a non-linear optimization algorithm;

[0084] S12: Use the ORB-SLAM3 algorithm to generate real-time map information based on the sensing data for dense mapping;

[0085] S13: Control the robot body to move to a preset position based on the optimized position information and the real-time map information;

[0086] S14: Control the robot body to collect the operation information of the belt conveyor to be detected, and send the operation information to the preset network structure to the control terminal.

[0087] It should be noted that Figure 2 The method provided corresponds to the robot in the above embodiments. Therefore, for the corresponding implementation solutions and beneficial effects, please refer to the part of the above robot, and no specific limitations will be made here.

[0088] In the above method, the type of the position sensor and the specific scheme for estimating the position information of the robot are not limited. Exemplarily, in some embodiments of the present application, the position sensor includes a lidar, an odometer, and an IMU. Figure 3 It is a flowchart of a method for obtaining position information provided by an embodiment of the present application. As Figure 3 shown, using the Cartographer algorithm to obtain the sensing data of the position sensor, and using a Bayesian filter to estimate the position information of the robot based on the sensing data includes:

[0089] S10-1: Obtain the processed sensing data of the position sensor, and perform scan matching on the sensing data;

[0090] It should be noted that an inertial measurement unit (IMU) is mainly a sensor used to detect and measure acceleration and rotational motion. The sensing data in this embodiment includes the point cloud data of the lidar filtered by a voxel filter, the odometer sensing data screened by an attitude extrapolator, and the IMU data calibrated by a gravity alignment device. It can be understood that the corresponding collected data includes the position information of the robot (lidar - point cloud data), the moving distance of the robot (odometer data), and the inertial information of the robot (IMU data).

[0091] S10-2: Use a Bayesian filter to filter and screen the sensing data, and discard the sensing data that does not meet the preset requirements;

[0092] It should be noted that Bayesian filtering is a method of probabilistic inference used to estimate the state of a system over time. In this embodiment, the specific content of the preset requirements is not limited.

[0093] S10-3: Establish a position submap based on the screened sensing data and update the voxel network;

[0094] A voxel is short for Volume Pixel. The three-dimensional object containing voxels can be represented by volume rendering or by extracting the polygonal isosurface of a given threshold contour. As the name implies, it is the smallest unit for dividing digital data in three-dimensional space. In this embodiment, after establishing a corresponding subgraph based on the corresponding voxel network, the voxels are updated based on the filtered sensing network, so as to obtain the position contour modeling and volume rendering image of the corresponding robot.

[0095] S10-4: Insert the point cloud data filtered by the voxel filter into the voxel network to obtain the position information of the robot.

[0096] Finally, the filtered point cloud data is matched in the subgraph of the voxel network to obtain the corresponding position information of the robot. Generally, the above solutions are all completed by the front end of the corresponding central processor, that is, in this embodiment, a specific front end is given for obtaining position information, thereby further improving the integrity of the solution.

[0097] In the above embodiments, the optimization method for the specific non-linear optimization algorithm is not limited. In some embodiments of the present application, Figure 4 is a flowchart of a position information optimization method provided by an embodiment of the present application. As Figure 4 shown, using a non-linear optimization algorithm to optimize the position information includes:

[0098] S11-1: Use a non-linear optimization algorithm to calculate the intra-subgraph constraint and inter-subgraph constraint of the position subgraph to obtain the error of the position information;

[0099] It should be noted that when using a non-linear optimization algorithm to perform the constraint between subgraphs to obtain the corresponding error content, the error content here refers to the error information of the corresponding position information of the robot within the subgraph.

[0100] S11-2: Based on the odometry sensing data, IMU data, and pose data of the fixed frame, establish an optimization model;

[0101] It should be noted that the main reasons for the error are the movement distance and the deviation of the specific inertia. Therefore, the corresponding odometry sensing data, IMU data, and pose data of the fixed frame are comprehensively used to establish the corresponding optimization model. It can be understood that the pose data refers to the position and pose of the corresponding robot.

[0102] S11-3: Based on the optimization model, perform pose optimization on the error of the position information to obtain the optimized position information of the robot.

[0103] In this embodiment, a specific optimization scheme for the position of the robot is provided. An optimization model is established through odometry sensing data, IMU data, and pose data of the fixed frame, and corresponding error optimization is performed, thereby further improving the integrity of the scheme.

[0104] In the above embodiment, the specific method for generating a real-time map is not limited. Exemplarily, in some embodiments of the present application, Figure 5 is a flowchart of a method for generating a real-time map provided by an embodiment of the present application. As Figure 5 shown, using the ORB-SLAM3 algorithm to perform dense mapping based on the sensing data to generate real-time map information includes:

[0105] S12-1: Obtain the depth image information in the sensing data, and after preprocessing the sensing data, perform pose estimation on the depth image information and establish a local map;

[0106] It should be noted that the depth image information is the corresponding image content collected by the depth camera. After processing, some key information in the corresponding image can be obtained, and based on the corresponding key information, the position and pose are estimated, and the corresponding image-based map is established.

[0107] S12-2: Create a new key frame for the processed depth image information and insert the key frame into the thread of the local map;

[0108] It should be noted that the key frame is the image content of a certain frame extracted from the corresponding image information. By inserting the corresponding key frame into the corresponding map for comparison, the real-time map is updated and the content is screened. It should be noted that in this embodiment, the number of newly created key frames, etc. are not specifically limited.

[0109] S12-3: Perform three-dimensional bad point removal on the key frame and create new three-dimensional points for local BA optimization to remove redundant key frames;

[0110] Remove the three-dimensional bad points in the corresponding key frame. The three-dimensional bad points are the map points in the image that do not conform to the rules of the corresponding map or have problems. New three-dimensional points are created at the positions of the original three-dimensional bad points for compensation to optimize the corresponding three-dimensional bad points, and local BA optimization is used to remove redundant key frames, thereby further optimizing the corresponding local map.

[0111] S12-4: Establish a similarity model based on the processed key frame for loop closure fusion, and perform global optimization on the local map based on the similarity model to obtain real-time map information.

[0112] In this embodiment, the ORB-SLAM3 algorithm is adopted by VSLAM. This algorithm can achieve dense mapping and high-precision positioning, improving the robot's autonomous navigation and obstacle avoidance capabilities.

[0113] Considering the real-time update of the map information of the robot's moving position, exemplarily, in some embodiments of the present application, after obtaining the real-time map information by using the ORB-SLAM3 algorithm for dense mapping and high-precision positioning based on the sensing data, it further includes:

[0114] Extract the map data feature points from the sensing data and match them with the preset map;

[0115] Based on the matching result, detect whether the intelligent belt conveyor inspection robot has passed through the explored area and update the real-time map information.

[0116] In this embodiment, it is proposed that the front end part also includes some feature extraction and matching algorithms for extracting the feature points in the map and matching them with the previous map to facilitate the map optimization of the back end part. The back end part also includes some map update and loop closure detection algorithms for detecting whether the robot has passed through the area that has been explored before and updating the map information.

[0117] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. A belt conveyor intelligent inspection robot, characterized in that: include: A robot body and a position sensor, wherein the position sensor is arranged on the robot body, and the robot body is equipped with a central processor developed based on a robot operating system of the ROS2 framework; The robot body moves based on the control instruction of the central processor. The robot body is also used to collect the operation information of the belt conveyor to be detected, and send the operation information to the central processor for analysis to obtain the analysis result. The central processor sends the analysis result to the control terminal through a preset communication network. The robot operating system adopts the Cartographer algorithm to obtain the sensor data of the position sensor, uses a Bayesian filter to estimate the position information of the robot based on the sensor data, and uses a nonlinear optimization algorithm to optimize the position information. The robot operating system also adopts the ORB-SLAM3 algorithm to perform dense mapping based on the sensor data to generate real-time map information, and the central processor performs autonomous navigation of the robot based on the position information of the robot and the real-time map information.

2. The intelligent inspection robot for a belt conveyor according to claim 1, characterized in that: The robot body further includes a monitoring device, a chassis and a power system, wherein the chassis is arranged at the bottom of the shell of the robot body, the monitoring device is arranged on the shell, the signal output end of the monitoring device is connected to the signal receiving end of the central processor, the driving end of the power system is connected to the chassis, and the control end of the power system is connected to the central processor; The power system is used to receive control instructions from the central processor and drive the chassis to move based on the control instructions. The monitoring device is used to collect operating information of the belt conveyor to be detected and send the operating information to the central processor for analysis to obtain analysis results.

3. The intelligent inspection robot for a belt conveyor according to claim 2, characterized in that: Also includes: Suspension and shock absorbers, environmental sensors, and safety protection devices; The suspension and shock absorbing device is arranged on the chassis, the environmental sensor is used to monitor the gas concentration, pollutants, ambient temperature and humidity in the environment, the safety protection device includes a collision sensor and an emergency stop button, and the safety protection device is connected to the drive motor to perform emergency braking when an abnormal situation occurs in the robot.

4. The intelligent inspection robot for a belt conveyor according to claim 1, characterized in that: The preset communication network is a hierarchical network built based on a network communication protocol. The preset communication network includes a core network, a convergence network and an access network, and uses optical fiber or Gigabit Ethernet to access corresponding network communication equipment for communication.

5. A control method for a belt conveyor intelligent inspection robot, applied to the belt conveyor intelligent inspection robot as claimed in any one of claims 1 to 4, characterized in that: The steps include: The Cartographer algorithm is used to obtain sensor data of a position sensor, and a Bayesian filter is used to estimate the position information of the robot based on the sensor data; Optimizing the position information using a nonlinear optimization algorithm; Using the ORB-SLAM3 algorithm to generate real-time map information based on the sensor data for dense mapping; Controlling the robot body to move to a preset position based on the optimized position information and the real-time map information; The robot body is controlled to collect operation information of the belt conveyor to be detected, and the operation information is sent to the control terminal in a preset network structure.

6. A belt conveyor intelligent inspection robot control method as claimed in claim 5, characterized in that: The position sensor includes a laser radar, an odometer, and an IMU. The Cartographer algorithm is used to obtain the sensor data of the position sensor, and the Bayesian filter is used to estimate the position information of the robot based on the sensor data, including: Acquire processed sensor data of the position sensor and perform scan matching on the sensor data, wherein the sensor data includes point cloud data of a laser radar filtered by a voxel filter, odometer sensor data filtered by an attitude extrapolator, and IMU data calibrated by a gravity alignment device; Using a Bayesian filter to filter the sensor data and discard sensor data that does not meet preset requirements; Establishing a position submap based on the filtered sensor data and updating a voxel network; The point cloud data filtered by the voxel filter is inserted into the voxel network to obtain the position information of the robot.

7. A belt conveyor intelligent inspection robot control method as claimed in claim 6, characterized in that: Optimizing the position information using a nonlinear optimization algorithm includes: Using a nonlinear optimization algorithm to calculate the intra-subgraph constraints and inter-subgraph constraints of the position subgraph to obtain the error of the position information; Establishing an optimization model based on the odometer sensor data, the IMU data and the position and posture data of the fixed frame; Based on the optimization model, the error of the position information is optimized to obtain the optimized position information of the robot.

8. The control method of a belt conveyor intelligent inspection robot according to claim 5, characterized in that: The method of using the ORB-SLAM3 algorithm to generate real-time map information by dense mapping based on the sensor data includes: Acquire the depth image information in the sensor data, and after preprocessing the sensor data, perform pose estimation on the depth image information and establish a local map; Creating a new key frame for the processed depth image information, and inserting the key frame into the thread of the local map; Eliminate three-dimensional bad pixels from the key frames, and create new three-dimensional points for local BA optimization to eliminate redundant key frames; A similarity model is established based on the processed key frames to perform closed-loop fusion, and the local map is globally optimized based on the similarity model to obtain the real-time map information.

9. A belt conveyor intelligent inspection robot control method as claimed in claim 5, characterized in that: After the ORB-SLAM3 algorithm is used to perform dense mapping and high-precision positioning based on the sensor data to obtain real-time map information, the method further includes: Extracting map data feature points from the sensor data, and matching the map with a preset map based on the feature points; Based on the matching results, it is detected whether the belt conveyor intelligent inspection robot has passed through the explored area, and the real-time map information is updated.

Images