Intelligent detection robot of a sampling mechanism

Abstract

The application provides an intelligent detection robot of a novel sampling mechanism, which comprises a machine body, a moving mechanism, a sampling mechanism, a navigation positioning module, a detection module and a control system; the navigation positioning module is internally integrated with a vehicle-mounted positioning indicator lamp and a positioning navigation algorithm, and is used for high-precision positioning in a dark environment; the moving mechanism is symmetrically arranged on both sides of the machine body; the sampling mechanism and the detection module are arranged on the front side of the machine body; the navigation positioning module and the control system are arranged on the top of the machine body; and the moving mechanism, the sampling mechanism, the navigation positioning module, the detection module and the control system are electrically connected. The robot designed by the application realizes full-process automation of sampling, moving, positioning and detection, does not need manual intervention, improves work efficiency, reduces manual operation strength and human error, and is suitable for large-scale and multi-batch material sampling and detection scenes.

Description

Technical Field

[0001] This invention relates to the field of intelligent testing equipment technology, and in particular to an intelligent testing robot for a sampling mechanism. Background Technology

[0002] In the field of material quality testing, sampling is a crucial step in obtaining test samples, directly affecting the accuracy and representativeness of the test results. Currently, sampling equipment on the market is mainly divided into two categories: manual operation and semi-automatic operation, both of which have several technical shortcomings.

[0003] Firstly, existing sampling mechanisms mostly rely on power sources such as motors and cylinders to drive the material handling components, which not only increases the overall weight, size, and energy consumption of the equipment, but also poses a risk of operation interruption due to power source failure. This limits their applicability, especially in remote storage areas or scenarios without external power supply. At the same time, the presence of a power source also increases the manufacturing cost and maintenance difficulty of the equipment.

[0004] Secondly, the moving mechanism of traditional sampling equipment mostly adopts a wheeled structure. In scenarios such as warehouses and grain depots, the ground may have scattered materials, uneven surfaces, and a lot of dust. The wheeled structure is prone to slipping and jamming, resulting in poor equipment movement stability and inability to flexibly reach the designated sampling position, thus affecting work efficiency.

[0005] Third, some sampling scenarios (such as warehouses, underground storage rooms, and enclosed silos where nighttime operations are conducted) are in environments with insufficient lighting or complete darkness. The navigation and positioning of existing sampling equipment mostly rely on visual navigation or ordinary GPS positioning. Visual navigation fails in dark environments, and ordinary GPS has weak signals indoors or in enclosed spaces, making it impossible to achieve high-precision positioning. This results in large deviations in sampling positions, insufficient sample representativeness, and consequently affects the reliability of the test results.

[0006] To address the aforementioned issues, there is an urgent need to develop an intelligent sampling and testing device with a power-free material handling structure, stable mobility, and adaptability to dark environments, in order to improve the automation level, stability, and accuracy of sampling operations. Summary of the Invention

[0007] In view of the above, the main objective of this invention is to provide an intelligent inspection robot for a sampling mechanism to solve the aforementioned technical problems.

[0008] This invention proposes an intelligent inspection robot for a sampling mechanism, comprising a body, a moving mechanism, a sampling mechanism, a navigation and positioning module, a detection module, and a control system; the navigation and positioning module integrates a vehicle-mounted positioning indicator and a positioning and navigation algorithm, and is used for high-precision positioning in dark environments;

[0009] Furthermore, the moving mechanism is symmetrically arranged on both sides of the machine body, the sampling mechanism and the detection module are both located on the front side of the machine body, and the navigation and positioning module and the control system are both located on the top of the machine body; the moving mechanism, the sampling mechanism, the navigation and positioning module, and the detection module are all electrically connected to the control system.

[0010] Furthermore, the moving mechanism adopts a tracked structure, which includes a drive wheel, a track, a guide wheel, and a track tensioning device. The drive wheels are symmetrically arranged on both sides of the machine body, electrically connected to the control system, and used to drive the track to rotate.

[0011] Furthermore, the track is fitted onto the drive wheel and guide wheel, and the track is made of highly elastic and wear-resistant rubber material with anti-slip protrusions on the surface; the guide wheel is evenly distributed on the inner side of the track and is used to support the track; the track tensioning device is symmetrically arranged on both sides of the machine body and is used to adjust the track tension to avoid the track being too loose or too tight.

[0012] Furthermore, the sampling mechanism includes a sampling rod, an electric push rod, a drill bit, a material inlet, and a triggering component; the detection module is located below the material inlet and is used to perform real-time detection on the sample in the material inlet, and the detection results are wirelessly transmitted to the back-end terminal through the control system; the number of vehicle positioning indicator lights is four, and the four vehicle positioning indicator lights are arranged around the four corners of the top of the machine body.

[0013] Furthermore, the electric push rod is fixedly installed on one side of the top of the machine body, and the sampling rod is fixedly installed on the output end of the bottom of the electric push rod, and the sampling rod has a hollow structure; the drill bit is set at the bottom of the sampling rod, and the material inlet is set on the side wall in the middle of the sampling rod; a temperature and humidity sensor is also set at the bottom of the side wall of the sampling rod, and the temperature and humidity sensor is located above the drill bit, and the temperature and humidity sensor is used to detect the temperature and humidity of the sample.

[0014] Furthermore, the triggering component includes a sampling rod housing and a spring. The sampling rod housing is movably sleeved on the outside of the sampling rod and can move up and down on the outside of the sampling rod. A discharge port is also provided in the middle of the sampling rod housing. The spring is sleeved on the side wall of the sampling rod and is located between the sampling rod housing and the sampling rod.

[0015] Furthermore, a detection tray and a feeding lever are also installed on the front side of the machine body. The detection tray is located at the bottom of the sampling rod, and the feeding lever is located above the detection tray and on one side of the sampling rod.

[0016] Furthermore, the positioning and navigation algorithm specifically includes the following steps:

[0017] Step 1: Use multiple infrared cameras pre-positioned on the top of the work area to perform multi-view synchronous scanning of the work area, collect environmental feature point cloud data to obtain environmental image information; construct a three-dimensional environmental map containing the outline and texture features of fixed structures based on the environmental image information, and input the three-dimensional environmental map into the control system.

[0018] The spatial distance between the four vehicle positioning indicator lights is measured using an optical rangefinder to obtain the initial distance between the four vehicle positioning indicator lights; the initial distance between the four vehicle positioning indicator lights is used as the reference spatial relationship data.

[0019] Step 2: After the robot starts, use an infrared camera to continuously acquire optical images including the vehicle positioning indicator light; perform Gaussian filtering and histogram equalization on each frame of optical image to obtain the preprocessed image;

[0020] The preprocessed image is converted from RGB color space to HSV color space. Then, the light spot areas of four vehicle positioning indicator lights are extracted according to the preset brightness and saturation thresholds. The halo interference is eliminated by morphological closing operation to obtain four preprocessed light spot areas.

[0021] Contour extraction and ellipse fitting were performed on the four preprocessed spot regions respectively, and the centroid pixel coordinates of the four vehicle positioning indicator lights were calculated.

[0022] Based on the centroid pixel coordinates of the four vehicle positioning indicator lights, the pixel distance between each vehicle positioning indicator light in the preprocessed image is calculated. Then, combined with the camera calibration parameters, the pixel distance is converted into the actual physical distance to obtain real-time measured distance data.

[0023] Step 3: Subtract the real-time measured distance data from the reference spatial relationship data to obtain the distance difference score; establish a spatial geometric deformation model of the four vehicle-mounted positioning indicator lights under robot posture changes;

[0024] The distance difference value is input into the spatial geometric deformation model, and the least squares optimization algorithm is used to solve for the combination of attitude angles that minimizes the geometric deformation error. A Kalman filter is then used for smoothing to obtain the robot's attitude angle data. The robot's attitude angle data includes pitch angle, roll angle, and orientation angle.

[0025] Step 4: Perform steps 2 and 3 independently on the multiple infrared cameras to obtain multiple sets of real-time distance measurement data and multiple sets of robot attitude angle data;

[0026] The robot's attitude angle data is used as the initial attitude estimate. Based on the installation position and extrinsic matrix of each infrared camera, multiple sets of real-time measured distance data are uniformly transformed into the pre-constructed global environment map coordinate system to obtain multiple sets of standard coordinate data.

[0027] An adaptive weighted fusion algorithm based on covariance is used to fuse multiple sets of standard coordinate data to obtain the optimal estimated position. The optimal estimated position is then matched with the 3D environment map by feature matching, and fine alignment is performed by an iterative nearest point algorithm to obtain the robot's pose information in the 3D environment map.

[0028] Step 5: Based on the robot's pose information in the 3D environment map and the preset target coordinates of the sampling points, construct a 2D occupancy grid map in the 3D environment map.

[0029] An improved A* heuristic search algorithm is used for path planning: the Manhattan distance and the turning penalty term are used together to form a heuristic function to search for the collision-free optimal path from the current position to the target point in a two-dimensional occupied grid map, and the path nodes are smoothed by B-spline curves to obtain the final planned path.

[0030] The control system controls the robot to move along the final planned path. During the movement, steps 2 to 4 are executed cyclically at a fixed frequency to achieve real-time pose update and positioning closed loop.

[0031] The robot's real-time pose information in the 3D environment map is compared with the planned path to generate lateral and longitudinal tracking errors. The robot's speed and steering are dynamically adjusted by a PID controller until the robot accurately reaches the target sampling point, completing the fully autonomous, high-precision navigation and positioning task in the dark environment.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] 1. The sampling mechanism designed in this invention adopts a non-power source material picking structure, which eliminates the traditional power source such as cylinders, significantly reduces the weight, volume and energy consumption of the equipment, reduces the failure points, and improves the applicability of the equipment in scenarios without external power supply; at the same time, it simplifies the structural design, reduces manufacturing costs and maintenance difficulty, and the material picking process is achieved by mechanical force and elastic reset force, so the operation is stable and reliable.

[0034] 2. The mobile mechanism designed in this invention adopts a tracked structure, which has stronger ground adaptability compared to the traditional wheeled chassis. It can move stably on uneven ground, where materials are scattered, or where there is a lot of dust, avoiding slippage and jamming. The design of high-elasticity wear-resistant tracks and anti-slip protrusions further improves the equipment's mobility and movement stability, ensuring that the robot can accurately reach the designated sampling position.

[0035] 3. This invention utilizes a collaborative design of "infrared camera + vehicle-mounted positioning indicator light + positioning and navigation algorithm". The infrared camera can accurately capture images of the positioning lights in dark environments. The four vehicle-mounted positioning indicator lights are arranged in a fixed rectangle and provide a preset initial distance reference. The positioning and navigation algorithm calculates the real-time distance between the vehicle-mounted positioning indicator lights in the image and compares it with the preset initial distance to accurately deduce the robot's orientation angle, pitch angle, and roll angle. This enables high-precision positioning in dark environments, ensuring the accuracy of sampling positions and improving the representativeness of samples and the reliability of detection results.

[0036] 4. The robot designed in this invention realizes full automation of sampling, movement, positioning and detection without human intervention, which improves work efficiency, reduces the intensity of manual operation and human error, and is suitable for large-scale, multi-batch material sampling and detection scenarios.

[0037] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by means of embodiments of the invention. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the overall structure of the present invention;

[0039] Figure 2 This is the front view of the present invention;

[0040] Figure 3 This is a schematic diagram of the sampling mechanism of the present invention.

[0041] In the diagram, 1. machine body, 2. sampling rod, 3. electric push rod, 4. drive wheel, 5. track, 6. guide wheel, 7. detection tray, 8. vehicle positioning indicator light, 9. detection module, 10. drill bit, 11. spring, 12. sampling rod housing, 13. temperature and humidity sensor, 14. material inlet, 15. control system, 16. material feeding lever, 17. material outlet. Detailed Implementation

[0042] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0043] These and other aspects of the embodiments of the present invention will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific embodiments of the present invention are specifically disclosed to illustrate some ways of implementing the principles of the embodiments of the present invention; however, it should be understood that the scope of the embodiments of the present invention is not limited thereto.

[0044] Please see Figures 1 to 3 This embodiment provides an intelligent inspection robot for a sampling mechanism, including a body 1, a moving mechanism, a sampling mechanism, a navigation and positioning module, a detection module 9, and a control system 15. The navigation and positioning module integrates a vehicle positioning indicator light 8 and a positioning and navigation algorithm, and is used for high-precision positioning in dark environments. The moving mechanism is symmetrically arranged on both sides of the body 1, the sampling mechanism and the detection module 9 are both located on the front side of the body 1, and the navigation and positioning module and the control system 15 are both located on the top of the body 1. The moving mechanism, the sampling mechanism, the navigation and positioning module, and the detection module 9 are all electrically connected to the control system 15.

[0045] The moving mechanism adopts an integral track 5 structure, which includes a drive wheel 4, a track 5, a guide wheel 6, and a track tensioning device. The drive wheel 4 is symmetrically arranged on both sides of the machine body 1. The drive wheel 4 is electrically connected to the control system 15 and is used to drive the track 5 to rotate. The track 5 is fitted on the drive wheel 4 and the guide wheel 6. The track 5 is made of high-elasticity wear-resistant rubber material and has anti-slip protrusions on the surface. The guide wheel 6 is evenly distributed on the inner side of the track 5 and is used to support the track 5. The track tensioning device is symmetrically arranged on both sides of the machine body 1 and is used to adjust the tension of the track 5 to prevent the track 5 from being too loose or too tight.

[0046] In a specific embodiment, the tracked structure of the moving mechanism is specifically configured as follows: the track 5 is made of high-elasticity nitrile rubber, with a width of 150mm, and the surface is provided with evenly distributed diamond-shaped anti-slip protrusions; the drive wheel 4 adopts a gear structure and is connected to a stepper motor, which is controlled by a control system; the guide wheel 6 is made of polyurethane and is evenly distributed on the track support frame inside the track 5; the track tensioning device adopts a screw adjustment structure, and the tension of the track 5 is adjusted by rotating the screw to push the guide wheel 6 to move.

[0047] The sampling mechanism adopts a non-powered material handling structure. The sampling mechanism includes a sampling rod 2, an electric push rod 3, a drill bit 10, a material handling port 14, and a triggering component. The detection module 9 is located below the material handling port 14 and is used to perform real-time detection on the sample in the material handling port 14. The detection results are wirelessly transmitted to the back-end terminal through the control system 15. There are four vehicle positioning indicator lights 8, which are arranged around the four corners of the top of the machine body 1. The electric push rod 3 is fixedly installed on one side of the top of the machine body 1, and the sampling rod 2 is fixedly installed on the output end at the bottom of the electric push rod 3. The sampling rod 2 has a hollow structure. The drill bit 10 is located at the bottom of the sampling rod 2, and the material handling port 14 is located on the side wall in the middle of the sampling rod 2. A temperature and humidity sensor 13 is also provided at the bottom of the side wall of the sampling rod 2, and the temperature and humidity sensor 13 is located above the drill bit 10. The temperature and humidity sensor 13 is used to detect the temperature and humidity of the sample.

[0048] The triggering assembly includes a sampling rod housing 12 and a spring 11. The sampling rod housing 12 is movably sleeved on the outside of the sampling rod 2, and the sampling rod housing 12 can move up and down on the outside of the sampling rod 2. A discharge port 17 is also provided in the middle of the sampling rod housing 12. The spring 11 is sleeved on the side wall of the sampling rod 2, and the spring 11 is located between the sampling rod housing 12 and the sampling rod 2. A detection tray 7 and a feeding paddle 16 are also installed on the front side of the machine body 1. The detection tray 7 is located at the bottom of the sampling rod 2, and the feeding paddle 16 is located above the detection tray 7 and is located on one side of the sampling rod 2.

[0049] In a specific embodiment: the sampling rod 2 is a hollow structure with a diameter of 34mm and a length of 250mm. A feeding port 14 is provided on the side wall of the middle part of the sampling rod 2. The spring 11 is a cylindrical helical spring, which is sleeved inside the sampling rod 2. One end of the spring abuts against the shoulder of the outer shell 12 of the sampling rod, and the other end abuts against the mounting seat of the sampling rod 2. The outer shell 12 of the sampling rod moves upward under the resistance of the grain, thereby opening the feeding port 14. When resetting, the outer shell 12 of the sampling rod will close the feeding port 14 under the action of gravity and spring force.

[0050] The detection module 9 uses a camera with an IMX258 main controller, which is set directly above the detection tray 7. After the sampling mechanism completes the material collection, the sample falls into the detection area of ​​the detection module 9. The sensor detects indicators such as mold and grain insect content of the sample, and the detection results are wirelessly transmitted to the back-end terminal through the control system 15.

[0051] When the robot moves to the sampling position, the control system 15 stops the moving mechanism and then controls the electric push rod 3 to push the sampling rod 2 towards the material. The material exerts a reverse force on the sampling rod housing 12, and the sampling rod housing 12 moves upward after encountering resistance, exposing the material intake port 14 on the side of the sampling rod 2. When the sampling rod 2 is inserted into the material to a preset depth, the material will enter the interior of the material intake port 14 from the material intake port 14. After the material intake is completed, the control system controls the electric push rod 3 to drive the sampling rod 2 to move in the opposite direction. The sampling rod housing 12 moves downward under the action of gravity and spring force, thereby closing the material intake port 14 on the side of the sampling rod 2, completing the material intake. The material intake process is achieved by mechanical force and elastic restoring force, and the action is stable and reliable.

[0052] In a preferred embodiment of the present invention, the positioning and navigation algorithm specifically includes the following steps:

[0053] Step 1: Use four infrared cameras pre-positioned on the top of the work area to perform multi-view synchronous scanning of the work area, collect environmental feature point cloud data to obtain environmental image information; construct a three-dimensional environmental map containing the outline and texture features of fixed structures based on the environmental image information, and input the three-dimensional environmental map into the control system.

[0054] The spatial distance between the four vehicle positioning indicator lights is measured using an optical distance measuring instrument to obtain the initial distance between the four vehicle positioning indicator lights; the initial distance between the four vehicle positioning indicator lights is used as the reference spatial relationship data.

[0055] In step 1, the four infrared cameras are Hikvision MV-CA016-10GC models, which have infrared night vision function and an imaging resolution of 4024×3036. They are installed at the front, rear and left and right sides of the grain silo to ensure that there are no blind spots in the camera's shooting angle.

[0056] Step 2: After the robot starts, use an infrared camera to continuously acquire optical images including the vehicle positioning indicator light; perform Gaussian filtering and histogram equalization on each frame of optical image to obtain the preprocessed image;

[0057] The preprocessed image is converted from RGB color space to HSV color space. Then, the light spot areas of four vehicle positioning indicator lights are extracted according to the preset brightness and saturation thresholds. The halo interference is eliminated by morphological closing operation to obtain four preprocessed light spot areas.

[0058] Contour extraction and ellipse fitting were performed on the four preprocessed spot regions respectively, and the centroid pixel coordinates of the four vehicle positioning indicator lights were calculated.

[0059] Based on the centroid pixel coordinates of the four vehicle positioning indicator lights, the pixel distance between each vehicle positioning indicator light in the preprocessed image is calculated. Then, combined with the camera calibration parameters, the pixel distance is converted into the actual physical distance to obtain real-time measurement distance data.

[0060] Step 3: Subtract the real-time measured distance data from the reference spatial relationship data to obtain the distance difference score; establish a spatial geometric deformation model of the four vehicle-mounted positioning indicator lights under robot posture changes;

[0061] The distance difference value is input into the spatial geometric deformation model, and the least squares optimization algorithm is used to solve for the combination of attitude angles that minimizes the geometric deformation error. A Kalman filter is then used for smoothing to obtain the robot's attitude angle data. The robot's attitude angle data includes pitch angle, roll angle, and orientation angle.

[0062] It should be noted that, given the theoretical three-dimensional coordinates of the pitch angle, roll angle, orientation angle, and four vehicle positioning indicator lights, the spatial geometric deformation model calculates the theoretical distance between any two vehicle positioning indicators based on these theoretical coordinates. This constructs an error function with the attitude angle as the optimization variable. The function value of this error function is the sum of the squares of the differences between all corresponding theoretical distances and real-time measured distances. The Levenburg-Marquardt nonlinear least squares algorithm is then used to iteratively solve for the attitude angle that minimizes this error function, which is then used as the initial attitude angle solution.

[0063] Step 4: Perform steps 2 and 3 independently on the multiple infrared cameras to obtain multiple sets of real-time distance measurement data and multiple sets of robot attitude angle data;

[0064] The robot's attitude angle data is used as the initial attitude estimate. Based on the installation position and extrinsic matrix of each infrared camera, multiple sets of real-time measured distance data are uniformly transformed into the pre-constructed global environment map coordinate system to obtain multiple sets of standard coordinate data.

[0065] An adaptive weighted fusion algorithm based on covariance is used to fuse multiple sets of standard coordinate data to obtain the optimal estimated position. The optimal estimated position is then matched with the 3D environment map by feature matching, and fine alignment is performed by an iterative nearest point algorithm to obtain the robot's pose information in the 3D environment map.

[0066] It should be noted that the specific steps of using the covariance-based adaptive weighted fusion algorithm to fuse multiple sets of standard coordinate data include: calculating the covariance matrix of the centroid pixel coordinates of the vehicle positioning indicator obtained from each infrared camera viewpoint to characterize the uncertainty of the observation; dynamically allocating fusion weights according to the magnitude of the inverse of the covariance matrix (i.e., the information matrix) of each viewpoint observation; and calculating the mean of all weighted coordinates as the optimal estimated position.

[0067] The precise alignment using the iterative nearest-point algorithm specifically includes: taking the coordinates of the four vehicle positioning indicator lights in the optimal estimated position as the source point cloud, and performing iterative nearest-point registration with the reference coordinates (target point cloud) of the same group of vehicle positioning indicator lights pre-labeled in the 3D environment map; calculating the optimal rigid transformation matrix (rotation matrix and translation vector) from the source point cloud to the target point cloud; and applying the optimal rigid transformation matrix to the robot's current position estimation to output the robot's pose information in the 3D environment map.

[0068] Step 5: Based on the robot's pose information in the 3D environment map and the preset target coordinates of the sampling points, construct a 2D occupancy grid map in the 3D environment map.

[0069] An improved A* heuristic search algorithm is used for path planning: the Manhattan distance and the turning penalty term are used together to form a heuristic function to search for the collision-free optimal path from the current position to the target point in a two-dimensional occupied grid map, and the path nodes are smoothed by B-spline curves to obtain the final planned path.

[0070] The control system controls the robot to move along the final planned path. During the movement, steps 2 to 4 are executed cyclically at a fixed frequency to achieve real-time pose update and positioning closed loop.

[0071] The robot's real-time pose information in the 3D environment map is compared with the planned path to generate lateral and longitudinal tracking errors. The robot's speed and steering are dynamically adjusted through a PID controller until the robot accurately reaches the target sampling point, completing the fully autonomous, high-precision navigation and positioning task in the dark environment.

[0072] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0073] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. An intelligent inspection robot for a sampling mechanism, characterized in that, It includes a fuselage, a moving mechanism, a sampling mechanism, a navigation and positioning module, a detection module, and a control system; the navigation and positioning module integrates a vehicle positioning indicator and a positioning and navigation algorithm, and is used for high-precision positioning in dark environments; The moving mechanism is symmetrically arranged on both sides of the machine body, the sampling mechanism and the detection module are both located on the front side of the machine body, and the navigation and positioning module and the control system are both located on the top of the machine body; the moving mechanism, the sampling mechanism, the navigation and positioning module, and the detection module are all electrically connected to the control system. The sampling mechanism includes a sampling rod, an electric push rod, a drill bit, a feeding port, and a triggering component; the detection module is located below the feeding port and is used to perform real-time detection on the sample in the feeding port, and the detection results are wirelessly transmitted to the back-end terminal through the control system; the number of vehicle positioning indicator lights is four, and the four vehicle positioning indicator lights are arranged around the four corners of the top of the machine body; The electric push rod is fixedly installed on one side of the top of the machine body, and the sampling rod is fixedly installed on the output end of the bottom of the electric push rod, and the sampling rod has a hollow structure; the drill bit is set at the bottom of the sampling rod, and the material inlet is set on the side wall in the middle of the sampling rod; a temperature and humidity sensor is also set at the bottom of the side wall of the sampling rod, and the temperature and humidity sensor is located above the drill bit. The temperature and humidity sensor is used to detect the temperature and humidity of the sample. The triggering component includes a sampling rod housing and a spring. The sampling rod housing is movably sleeved on the outside of the sampling rod and can move up and down on the outside of the sampling rod. A discharge port is also provided in the middle of the sampling rod housing. The spring is sleeved on the side wall of the sampling rod and is located between the sampling rod housing and the sampling rod. The machine body is equipped with a detection tray and a feeding lever on the front side. The detection tray is located at the bottom of the sampling rod, and the feeding lever is located above the detection tray and on one side of the sampling rod.

2. The intelligent inspection robot of the sampling mechanism according to claim 1, characterized in that, The moving mechanism adopts a tracked structure, which includes a drive wheel, a track, a guide wheel, and a track tensioning device. The drive wheels are symmetrically arranged on both sides of the machine body, and are electrically connected to the control system. The drive wheels are used to drive the track to rotate.

3. The intelligent inspection robot of the sampling mechanism according to claim 2, characterized in that, The track is fitted onto the drive wheel and guide wheel. The track is made of highly elastic and wear-resistant rubber with anti-slip protrusions on the surface. The guide wheels are evenly distributed on the inner side of the track and are used to support the track. The track tensioning device is symmetrically arranged on both sides of the machine body and is used to adjust the track tension to prevent the track from being too loose or too tight.

4. The intelligent inspection robot of the sampling mechanism according to claim 1, characterized in that, The positioning and navigation algorithm specifically includes the following steps: Step 1: Use multiple infrared cameras pre-positioned on the top of the work area to perform multi-view synchronous scanning of the work area, collect environmental feature point cloud data to obtain environmental image information; construct a three-dimensional environmental map containing the outline and texture features of fixed structures based on the environmental image information, and input the three-dimensional environmental map into the control system. The spatial distance between the four vehicle positioning indicator lights is measured using an optical rangefinder to obtain the initial distance between the four vehicle positioning indicator lights; the initial distance between the four vehicle positioning indicator lights is used as the reference spatial relationship data. Step 2: After the robot is started, use an infrared camera to continuously collect optical images including the vehicle positioning indicator light; Gaussian filtering and histogram equalization are performed on each frame of optical image to obtain the preprocessed image; The preprocessed image is converted from RGB color space to HSV color space. Then, the light spot areas of four vehicle positioning indicator lights are extracted according to the preset brightness and saturation thresholds. The halo interference is eliminated by morphological closing operation to obtain four preprocessed light spot areas. Contour extraction and ellipse fitting were performed on the four preprocessed spot regions respectively, and the centroid pixel coordinates of the four vehicle positioning indicator lights were calculated. Based on the centroid pixel coordinates of the four vehicle positioning indicator lights, the pixel distance between each vehicle positioning indicator light in the preprocessed image is calculated. Then, combined with the camera calibration parameters, the pixel distance is converted into the actual physical distance to obtain real-time measured distance data. Step 3: Subtract the real-time measured distance data from the reference spatial relationship data to obtain the distance difference score; Establish a spatial geometric deformation model of four vehicle-mounted positioning indicator lights under robot posture changes; The distance difference value is input into the spatial geometric deformation model. The least squares optimization algorithm is used to solve for the combination of attitude angles that minimizes the geometric deformation error. The Kalman filter is then used for smoothing to obtain the robot's attitude angle data. The robot's attitude angle data includes: pitch angle, roll angle, and orientation angle; Step 4: Perform steps 2 and 3 independently on the multiple infrared cameras to obtain multiple sets of real-time distance measurement data and multiple sets of robot attitude angle data; The robot's attitude angle data is used as the initial attitude estimate. Based on the installation position and extrinsic matrix of each infrared camera, multiple sets of real-time measured distance data are uniformly transformed into the pre-constructed global environment map coordinate system to obtain multiple sets of standard coordinate data. An adaptive weighted fusion algorithm based on covariance is used to fuse multiple sets of standard coordinate data to obtain the optimal estimated position. The optimal estimated position is then matched with the 3D environment map by feature matching, and fine alignment is performed by an iterative nearest point algorithm to obtain the robot's pose information in the 3D environment map. Step 5: Based on the robot's pose information in the 3D environment map and the preset target coordinates of the sampling points, construct a 2D occupancy grid map in the 3D environment map. An improved A* heuristic search algorithm is used for path planning: the Manhattan distance and the turning penalty term are used together to form a heuristic function to search for the collision-free optimal path from the current position to the target point in a two-dimensional occupied grid map, and the path nodes are smoothed by B-spline curves to obtain the final planned path. The control system controls the robot to move along the final planned path. During the movement, steps 2 to 4 are executed cyclically at a fixed frequency to achieve real-time pose update and positioning closed loop. The robot's real-time pose information in the 3D environment map is compared with the planned path to generate lateral and longitudinal tracking errors. The robot's speed and steering are dynamically adjusted by a PID controller until the robot accurately reaches the target sampling point, completing the fully autonomous, high-precision navigation and positioning task in the dark environment.

Images