Uranium tank truck positioning system based on laser radar holder
Through the lidar gimbal scanning of truck point cloud data, filtering, clustering and coordinate system transformation, the automatic loading of uranium tanks is realized, solving the problems of low efficiency and safety hazards of uranium tank trucks in the existing technology, and improving the Loading efficiency and safety.
Patent Information
- Application Number
- CN202411094508.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-10
- Publication Date
- 2024-12-06
AI Technical Summary
The loading process of uranium tanks requires manual operation, resulting in low efficiency and safety risks, and the prior art cannot realize the automatic loading of uranium tanks.
Using a positioning system based on lidar gimbal, filtering, clustering, plane fitting and coordinate system transformation are performed by scanning the truck and outputting point cloud data, and the position and attitude of the saddle are automatically identified and positioned, thereby realizing the automation of uranium tanks. Load the car.
It realizes automatic loading of uranium tanks, improves loading efficiency, reduces the safety hazards of manual intervention, and can stably locate the saddle when the truck tilt angle and front data are missing, adapting to trucks of different lengths and widths. .
Smart Images

Figure CN119087450A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and specifically to a uranium tank loading positioning system based on a laser radar gimbal. Background Art
[0002] For safety and radiation protection reasons, uranium compounds need to be stored and transported in special sealed cans (hereinafter referred to as uranium cans). The uranium can is cylindrical in shape with three circular protrusions on the surface, which can serve as a limit when loading. Figure 6 When transporting uranium cans by truck, the cans need to be placed on a saddle, such as Figure 7 and 8 As shown, it is convenient to fix the uranium tank.
[0003] At present, loading of uranium canisters is done manually. When loading, the uranium canister is hoisted by a sling, and the sling is moved slowly closer to the saddle manually. When the uranium canister and the saddle are about to touch, it is necessary to look around the truck and adjust the position of the uranium canister to avoid collision, and at the same time make the circular protrusion of the uranium canister stuck outside the saddle. Therefore, the entire loading and unloading process takes a long time, is inefficient, and manual operation poses safety hazards. Summary of the invention
[0004] The purpose of this application is to provide a uranium tank loading positioning system based on a laser radar gimbal to solve the technical problems raised in the above background.
[0005] To achieve the above-mentioned purpose, the present application discloses the following technical solutions: a uranium tank loading positioning system based on a laser radar gimbal, comprising a laser radar gimbal and a positioning module communicating with the laser radar gimbal, wherein a parking area is provided at the bottom of the laser radar gimbal, and the positioning module is configured to: acquire the position and attitude of the loading spreader based on the truck in the parking area;
[0006] The acquiring of the position and posture of the loading spreader based on the truck in the parking area specifically includes:
[0007] S0: The truck is parked in the parking area, and the length direction of the truck is parallel to the X-axis of the preset O-XYZ coordinate system, and the width direction of the truck is parallel to the Y-axis of the preset O-XYZ coordinate system; the laser radar gimbal scans the truck and outputs the point cloud data converted to the O-XYZ coordinate system, and the point cloud data of the saddle is manually segmented and saved as a template point cloud; the sling is manually operated to place the uranium can in the middle of the saddle, and the coordinates and posture of the sling are recorded at this time;
[0008] S1: filtering the entire scene point cloud based on the coordinate range of the parking area to remove invalid point cloud data and ground point cloud;
[0009] S2: Perform Euclidean clustering on the remaining point cloud data, divide different points whose distance is greater than the first threshold into different point cloud clusters, and retain the largest point cloud cluster as the point cloud data of the truck, denoted as P;
[0010] S3: Calculate the normal of the point cloud P, and extract the angle between the normal and the Y axis that is less than the second threshold θ T The plane fitting is performed on the points of the truck, and the plane method is the random sampling consistency method to obtain the point cloud data of the truck side fender and construct the truck coordinate system O K -X K Y K Z K , where the normal of the plane is the inclination angle of the side baffle;
[0011] S4: Based on coordinate system O K -X K Y K Z K , correct the direction of the truck to be parallel to the O-XYZ coordinate system, calculate the bounding box of the truck point cloud, cut the bounding box according to the length and width data ratio of the vehicle body, remove the point cloud data of the front, rear, and both side fenders, and reversely transform the remaining bottom plate and saddle point cloud data back to the original posture;
[0012] S5: Calculate the normal of the point cloud and extract the angle between the normal and the z-axis that is less than the second threshold θ T The point cloud data of the truck bed is obtained by plane fitting based on the random sampling consistency normal, and all the point cloud data with a height greater than the third threshold from the bed plane are extracted as the point cloud data of the saddle;
[0013] S6: aligning the point cloud data of the saddle obtained in S5 with the template point cloud to obtain the transformation relationship between the point cloud data of the saddle and the template point cloud, and obtaining the position and posture of the spreader during current loading based on the spreader position and posture corresponding to the template point cloud.
[0014] Preferably, the S1 specifically includes:
[0015] For any point P in the point cloud i (x i ,y i ,z i ), retain the points that satisfy the filtering formula, and remove the points that do not satisfy the filtering formula; the filtering formula is specifically:
[0016]
[0017] Among them, x min is the minimum boundary value in the x direction, x max is the maximum boundary value in the x direction, y min is the minimum boundary value in the y direction, ymax is the maximum boundary value in the y direction, z min is the minimum boundary value in the Z direction, z max It is the maximum boundary value in the Z direction.
[0018] Preferably, in S3, the normal is calculated as follows:
[0019] Find the point p i The k points closest to each other are used to calculate the corresponding covariance matrix C. in, is the centroid of k points, and the eigenvector corresponding to the minimum eigenvalue of C is approximately expressed as point p i The normal n i =[x ni ,y ni ,z ni ].
[0020] Preferably, in S3, the angle between the extracted normal angle and the Y axis is less than a second threshold value θ T The plane fitting is performed on the points of the truck, and the plane method is the random sampling consistency method to obtain the point cloud data of the truck side fender, where the normal line of the plane is the inclination angle of the side fender, specifically including:
[0021] If i The angle between the Y axis and the T , then the point set Q is extracted;
[0022] The plane fitting based on random sampling consistency is performed on the point set Q to obtain the plane fitted by the point cloud data of the truck side fender. The unit normal vector of the obtained plane is expressed by n plane =[a K b K c K ] represents, and based on this, the truck coordinate system O is constructed K -X K Y K Z K ,in,
[0023] Preferably, in said S4, said based on the coordinate system O K -X K Y K Z K , correct the direction of the truck to be parallel to the O-XYZ coordinate system, including:
[0024] Based on the coordinate system O K -X K Y K Z K Construct rigid body transformation matrix T K :
[0025] R K =[X K Y K Z K ]
[0026]
[0027] For the point cloud P obtained by S2, transform P K =T K P, where the new point cloud P K The direction of the medium truck is corrected to be parallel to the coordinate axis, with the vehicle length direction parallel to the X axis and the vehicle width direction parallel to the Y axis.
[0028] Preferably, in S5, the normal of the calculated point cloud is calculated as the new point cloud P K The normal of .
[0029] Preferably, the S6 specifically includes:
[0030] Calculate the point cloud data P of the saddle separately an The centroid and template point cloud P m The centroid The point cloud data of the saddle and the template point cloud are offset by At the same time, construct the transformation matrix T 1 , T 1 for:
[0031]
[0032] Transform the point cloud data of the saddle to obtain a new point cloud Then, With the template point cloud P m The registration is performed by iterative nearest method. With P m Coincidentally, obtain the corresponding transformation matrix T 2 , at this time, the transformation matrix between the original saddle point cloud data and the template point cloud is:
[0033]
[0034] The position of the template point cloud is defined as η = [x η y η z η ] T , the grab angle is ω, then the current loading position and posture of the spreader are
[0035]
[0036] ωnow =ω-arctan2(r 4 ,r 1 )
[0037] Beneficial effects: Compared with the existing technical solutions, the uranium tank loading positioning system based on the laser radar gimbal of the present application can realize the automatic loading of uranium tanks. The process does not require human intervention, has high execution efficiency, and is safe and reliable. The present application can stably and efficiently identify and locate the position of the saddle on the truck, thereby providing accurate loading coordinates. In addition, it can also stably locate when the parking position of the truck has a certain inclination angle. The parking requirements for the truck are relatively loose and meet the actual use requirements. Moreover, the present application can adapt to different truck lengths and widths. In addition, when the front data is missing, it can also perform adaptive segmentation to accurately locate the coordinates of the saddle. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.
[0039] Figure 1 A schematic diagram of the structure of a uranium tank loading positioning system based on a laser radar pan-tilt provided in an embodiment of the present application;
[0040] Figure 2 A schematic diagram of the process of making a template point cloud portion provided in an embodiment of the present application;
[0041] Figure 3 A schematic diagram of a truck coordinate system provided in an embodiment of the present application;
[0042] Figure 4 A schematic diagram of the algorithm flow of the positioning part provided in the embodiment of the present application;
[0043] Figure 5 A schematic diagram of a process flow of a uranium tank loading processing algorithm provided in an embodiment of the present application;
[0044] Figure 6 This is a schematic diagram of the structure of a uranium tank;
[0045] Figure 7 This is a schematic diagram of the structure of the saddle;
[0046] Figure 8 This is a diagram of the saddle and truck after assembly. DETAILED DESCRIPTION
[0047] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.
[0048] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0049] This embodiment discloses a Figure 1 The uranium tank loading positioning system based on the laser radar gimbal shown (wherein the positioning module of the software part is not shown) comprises a laser radar gimbal and a positioning module communicating with the laser radar gimbal. The hardware consists of a set of laser radar gimbals, which are a rotating platform equipped with a laser radar. The rotation of the platform itself drives the laser radar to scan, thereby obtaining three-dimensional data of the scene. The laser radar gimbal is installed on the beam or the driving vehicle. After the empty truck is parked in the parking area, the laser radar gimbal is started to obtain the entire scene point cloud, and then a series of algorithm processing is performed on the scene point cloud to finally locate the loading position and posture of the saddle. The positioning module is configured to obtain the position and posture of the loading sling based on the truck in the parking area.
[0050] Specific, combined Figure 1-5 As shown, the processing algorithm for loading uranium tanks is based on the specific process of obtaining the position and posture of the loading hanger based on the truck in the parking area as follows:
[0051] S0: Calibration phase: Park a truck in a parking area in a standardized manner, and try to ensure that it is not tilted, that is, the length of the truck is parallel to the X-axis, and the width of the truck is parallel to the Y-axis. Scan the point cloud data of the empty truck. The point cloud data output by the sensor has been converted to the O-XYZ coordinate system. Manually segment the point cloud of the saddle and save it. This point cloud is the template point cloud. Manually operate the spreader to place the uranium can in the middle of the saddle, and record the coordinates and posture of the spreader at this time. Among them, the acquisition of the coordinates and posture of the spreader only needs to be operated once in the calibration phase, and this operation is not required in the subsequent operation process.
[0052] S1: Filter the entire scene point cloud based on the coordinate range of the parking area, remove most of the invalid point cloud data, and remove the ground point cloud. The specific operation is: for any point P in the point cloudi (x i ,y i ,z i ), the points that meet the filtering formula are retained, and the points that do not meet the conditions are eliminated. The specific filtering formula is:
[0053]
[0054] Among them, the boundary values of the three axes in the filtering formula (x min : The minimum boundary value in the x direction, x max : Maximum boundary value in the x direction, y min ,y max 、z min and z max Similarly) needs to be determined manually according to the actual situation, usually divided according to the parking area. It only needs to ensure that the entire parking area is within the boundary and the ground point cloud is removed.
[0055] S2: Perform Euclidean clustering on the remaining point clouds, divide different points whose distance is greater than a certain threshold into different point cloud clusters, and retain the largest point cloud cluster as the point cloud of the truck, denoted as P.
[0056] S3. Calculate the normal of point cloud P and extract the angle between the normal and the Y axis which is less than the threshold θ T (Usually set θ T =15°), plane fitting is performed, and the plane method is the random sampling consistency method, and the point cloud data of the truck side fender can be obtained. The normal line of the plane is the inclination angle of the side fender.
[0057] The normal is calculated by finding the point p i The k points closest to each other are used to calculate the corresponding covariance matrix.
[0058]
[0059] in, is the centroid of k points, and the eigenvector corresponding to the minimum eigenvalue of C can be approximately expressed as point p i The normal n i =[x ni ,y ni ,z ni ], if n i The angle between the Y axis and the T , then the point set Q is extracted.
[0060] Then, the point set Q is fitted with a plane based on the consistency of random sampling to obtain the plane fitted by the point cloud data of the truck side fender. The unit normal vector of the obtained plane is expressed as n plane =[a K bK c K ] represents, and based on this, the truck coordinate system O is constructed K -X K Y K Z K .
[0061] Y K =[a K b K c K ] T
[0062] X K =Y K ×Z
[0063] Z K =X K ×Y K
[0064] S4: Based on coordinate system O K -X K Y K Z K , correct the direction of the truck to be parallel to the O-XYZ coordinate system, and then calculate the bounding box of the truck point cloud, cut the bounding box according to the length and width data ratio of the vehicle body, remove the point cloud data of the front, rear, and both side fenders, and reversely transform the remaining bottom plate and saddle point cloud data back to the original posture.
[0065] First, based on the coordinate system O K -X K Y K Z K Construct rigid body transformation matrix T K :
[0066] R K =[X K Y K Z K ]
[0067]
[0068] For the point cloud P obtained by S2, transform P K =T K P, the new point cloud P K The direction of the truck has been corrected to be parallel to the coordinate axis (the length of the truck is parallel to the X axis, and the width of the truck is parallel to the Y axis).
[0069] S5: Calculate the normal of the point cloud and extract the angle between the normal and the z-axis that is less than the threshold θ T (Usually set θ T=15°), the point cloud data of the truck bed can be obtained by plane fitting with the random sampling consistency method, and all point cloud data with a height greater than a certain threshold from the bed plane are extracted, which is the point cloud data of the saddle.
[0070] S6: The saddle point cloud obtained in S5 is registered with the template point cloud to obtain the transformation relationship between the saddle point cloud and the template point cloud. Since the position and posture of the spreader corresponding to the template point cloud are known, the position and posture of the spreader during the current loading can be obtained.
[0071] Calculate the saddle point cloud P separately an The centroid and template point cloud P m The centroid Offset both the saddle point cloud and the template point cloud by At the same time, construct the transformation matrix T 1 for:
[0072]
[0073] Transform the saddle point cloud to get a new point cloud Then, With the template point cloud P m The registration is performed by iterative nearest method, so that With P m coincident, and the corresponding transformation matrix T 2 , then the transformation matrix between the original saddle point cloud and the template point cloud is
[0074]
[0075] The position of the template point cloud is η = [x η y η z η ] T , the grasping angle is ω, then the current loading position and posture are
[0076]
[0077] ω now =ω-arctan2(r 4 ,r 1 )
[0078] It should be noted that the acquisition of the overall scene point cloud in this solution is completed by fixed-point scanning with a laser radar gimbal, that is, the laser radar gimbal is installed in a fixed position for scanning, or installed on a vehicle, but the vehicle remains stationary during scanning, and the gimbal swings for scanning. It is also possible to not use a gimbal and directly install the laser radar on the vehicle, and use the linear motion of the vehicle to drive the laser radar to scan, but this method is less time-consuming and less stable than this solution.
[0079] Based on the above, it can be seen that the present application extracts the data of the saddle on the truck through the processing of the point cloud by the algorithm, and locates the position and posture of the saddle, providing accurate coordinates and posture for uranium tank loading.
[0080] In the embodiments provided in the present application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: application specific integrated circuit (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein or their combination. For software implementation, part or all of the flow of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication medium includes any medium that is convenient for transmitting a computer program from one place to another. The storage medium can be any available medium that a computer can access. The computer-readable storage medium can include but is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer.
[0081] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
Claims
1. A uranium tank loading positioning system based on laser radar gimbal, It is characterized in that The invention comprises a laser radar gimbal and a positioning module communicating with the laser radar gimbal, wherein a parking area is arranged at the bottom of the laser radar gimbal, and the positioning module is configured to obtain the position and attitude of the loading spreader based on the truck in the parking area; the position and attitude of the loading spreader based on the truck in the parking area is obtained, specifically including: S0: The truck is parked in the parking area, and the length direction of the truck is parallel to the X-axis of the preset O-XYZ coordinate system, and the width direction of the truck is parallel to the Y-axis of the preset O-XYZ coordinate system; the laser radar gimbal scans the truck and outputs the point cloud data converted to the O-XYZ coordinate system, and the point cloud data of the saddle is manually segmented and saved as a template point cloud; the sling is manually operated to place the uranium can in the middle of the saddle, and the coordinates and posture of the sling are recorded at this time; S1: filtering the entire scene point cloud based on the coordinate range of the parking area to remove invalid point cloud data and ground point cloud; S2: Perform Euclidean clustering on the remaining point cloud data, divide different points whose distance is greater than the first threshold into different point cloud clusters, and retain the largest point cloud cluster as the point cloud data of the truck, denoted as P; S3: Calculate the normal of the point cloud P, and extract the angle between the normal and the Y axis that is less than the second threshold θ T The plane fitting is performed on the points of the truck, and the plane method is the random sampling consistency method to obtain the point cloud data of the truck side fender and construct the truck coordinate system O K -X K Y K Z K , where the normal of the plane is the inclination angle of the side baffle; S4: Based on coordinate system O K -X K Y K Z K , correct the direction of the truck to be parallel to the O-XYZ coordinate system, calculate the bounding box of the truck point cloud, cut the bounding box according to the length and width data ratio of the vehicle body, remove the point cloud data of the front, rear, and both side fenders, and reversely transform the remaining bottom plate and saddle point cloud data back to the original posture; S5: Calculate the normal of the point cloud and extract the angle between the normal and the z-axis that is less than the second threshold θ T The point cloud data of the truck bed is obtained by plane fitting based on the random sampling consistency normal, and all the point cloud data with a height greater than the third threshold from the bed plane are extracted as the point cloud data of the saddle; S6: aligning the point cloud data of the saddle obtained in S5 with the template point cloud to obtain the transformation relationship between the point cloud data of the saddle and the template point cloud, and obtaining the position and posture of the spreader during current loading based on the spreader position and posture corresponding to the template point cloud.
2. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 1, It is characterized in that The S1 specifically includes: For any point P in the point cloud i (x i ,y i ,z i ), retain the points that satisfy the filtering formula, and remove the points that do not satisfy the filtering formula; the filtering formula is specifically: Among them, x min is the minimum boundary value in the x direction, x max is the maximum boundary value in the x direction, y min is the minimum boundary value in the y direction, y max is the maximum boundary value in the y direction, z min is the minimum boundary value in the Z direction, z max It is the maximum boundary value in the Z direction.
3. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 1, It is characterized in that In S3, the normal is calculated as follows: Find the point p i The k points closest to each other are used to calculate the corresponding covariance matrix C. in, is the centroid of k points, and the eigenvector corresponding to the minimum eigenvalue of C is approximately expressed as point p i The normal n i =[x ni ,y ni ,z ni ].
4. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 3, It is characterized in that In S3, the angle between the extracted normal line angle and the Y axis is less than the second threshold value θ T The plane fitting is performed on the points of the truck, and the plane method is the random sampling consistency method to obtain the point cloud data of the truck side fender, where the normal line of the plane is the inclination angle of the side fender, specifically including: If i The angle between the Y axis and the T , then the point set Q is extracted; The plane fitting based on random sampling consistency is performed on the point set Q to obtain the plane fitted by the point cloud data of the truck side fender. The unit normal vector of the obtained plane is expressed by n plane =[a K b K c K ] represents, and based on this, the truck coordinate system O is constructed K -X K Y K Z K ,in, 5. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 4, It is characterized in that In the S4, the coordinate system O K -X K Y K Z K , correct the direction of the truck to be parallel to the O-XYZ coordinate system, including: Based on the coordinate system O K -X K Y K Z K Construct rigid body transformation matrix T K : R K =[X K Y K Z K ] For the point cloud P obtained by S2, transform P K =T K P, where the new point cloud P K The direction of the medium truck is corrected to be parallel to the coordinate axis, with the vehicle length direction parallel to the X axis and the vehicle width direction parallel to the Y axis.
6. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 5, It is characterized in that In S5, the normal of the calculated point cloud is calculated as the new point cloud P K The normal of .
7. The uranium tank loading positioning system based on the laser radar pan-tilt according to claim 6, It is characterized in that The S6 specifically includes: Calculate the point cloud data P of the saddle separately an The centroid and template point cloud P m The centroid The point cloud data of the saddle and the template point cloud are offset by At the same time, construct the transformation matrix T 1 , T 1 for: Transform the point cloud data of the saddle to obtain a new point cloud Then, With the template point cloud P m The registration is performed by iterative nearest method. With P m Coincidentally, obtain the corresponding transformation matrix T 2 , at this time, the transformation matrix between the original saddle point cloud data and the template point cloud is: The position of the template point cloud is defined as η = [x η y η z η ] T , the grab angle is ω, then the current loading position and posture of the spreader are: ω now =ω-arctan2(r 4 ,r 1 )。
Citation Information
Cited By
Method, system and equipment for real-time measurement and calculation of hinge angle of automatic driving truck and medium
CN119270298A
Support device
CN121180749A
