CN109657698B - Magnetic suspension track obstacle detection method based on point cloud - Google Patents
Magnetic suspension track obstacle detection method based on point cloud Download PDFInfo
- Publication number
- CN109657698B CN109657698B CN201811386577.5A CN201811386577A CN109657698B CN 109657698 B CN109657698 B CN 109657698B CN 201811386577 A CN201811386577 A CN 201811386577A CN 109657698 B CN109657698 B CN 109657698B
- Authority
- CN
- China
- Prior art keywords
- point cloud
- obstacle
- magnetic suspension
- data
- points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention relates to a point cloud-based magnetic suspension track obstacle detection method, which comprises the following steps: 1) building a magnetic suspension track vehicle-mounted mobile scanning system; 2) acquiring original reference point cloud of a magnetic suspension rail surface and point cloud data to be detected through mobile scanning; 3) establishing a K-D tree index, and performing characteristic matching with point cloud data to be detected to obtain an obstacle point cloud; 4) after European clustering processing is carried out on the obstacle point cloud, an obstacle point cloud cluster set is obtained; 5) and finally, outputting the contour boundary and the position of each independent obstacle, thereby realizing the automatic detection of the obstacle on the magnetic suspension track surface. Compared with the prior art, the magnetic levitation system has the advantages of being strong in environment sensing capability, capable of playing an information supporting role in operation and maintenance safety of magnetic levitation and the like.
Description
Technical Field
The invention relates to a magnetic suspension track obstacle detection method, in particular to a magnetic suspension track obstacle detection method based on point cloud.
Background
Magnetic suspension traffic relies on electromagnetic force to make the train fly on the track in a 'ground-contacting' manner, has the advantages of high speed, low power consumption, no friction, low noise, long service life and the like, and is a novel advanced track traffic technology which is paid more and more attention by more and more countries. With the improvement of the requirement on the magnetic suspension design speed, the obstacle detection is carried out on the magnetic suspension track surface, and the obstacle is removed in time, so that the situation that the obstacle is electrified due to high-voltage current under the electrified condition can be avoided, and the track surface and the like are damaged to a certain extent; the danger caused by collision between the train and the barrier in the driving process can be avoided, and the safety of the driving route is ensured.
At present, few researches on the detection of obstacles on a magnetic suspension track surface are carried out, the researches mainly focus on the aspects of track gauge, smoothness and the like of a track, the magnetic suspension driving route is mainly checked in a manual mode, and the defects of high labor intensity, low efficiency and the like are overcome. Reference is made to a method for detecting obstacles on railway tracks, road traffic, etc. which includes: a stereo vision-based obstacle detection method, a lidar-based obstacle detection, a multi-sensor integration-based obstacle detection, an optical flow-based obstacle detection, and the like. (1) The barrier detection based on binocular stereo vision is a mainstream detection mode of the barrier at present, after left and right images of the barrier are obtained through a binocular camera, a target area of the barrier is obtained by adopting methods such as threshold segmentation and the like, and then characteristic point detection, extraction, stereo matching and the like are carried out to obtain the contour, position and depth information of the barrier, so that the barrier detection method has the advantages of strong flexibility, low cost and the like, but is greatly influenced by complex environments such as illumination, rain, snow and the like; (2) the detection method based on the laser radar emits laser through the laser emitter, then receives reflected light after the laser touches the surface of an object, measures the time difference between emission and reception to carry out distance measurement, can obtain various information such as the emission characteristic, the distance and the speed of a target, and finally detects obstacles through the partition, the clustering and the like of the obstacles and a road surface, and has the characteristics of simple principle, high precision, strong anti-interference performance and the like; (3) the detection method based on multi-sensor fusion includes that for example, a visual sensor is used for obtaining an obstacle image, a laser sensor, an infrared sensor or an ultrasonic sensor is used for describing an environment by combining a signal measured by the laser sensor, the infrared sensor or the ultrasonic sensor and the visual sensor after the approximate position of an obstacle is obtained by extracting the edge and the characteristics of the obstacle, the influence of environment change is small, the robustness is strong, the sensed information is reliable, the accuracy is high, the defect that information of a single sensor is insufficient is overcome, but the perfect fusion algorithm of different sensors is difficult and the cost is high; (4) the method is based on the optical flow method, a single camera is used for acquiring surrounding environment information at different moments, the optical flow is estimated through sequence images, and the optical flow field is segmented by using the characteristic that an obstacle in the optical flow is a sudden change, so that obstacle detection is realized.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a magnetic suspension track obstacle detection method based on point cloud.
The purpose of the invention can be realized by the following technical scheme:
a magnetic suspension track obstacle detection method based on point cloud comprises the following steps:
1) building a magnetic suspension track vehicle-mounted mobile scanning system;
2) acquiring original reference point cloud of a magnetic suspension rail surface and point cloud data to be detected through mobile scanning;
3) establishing a K-D tree index, and performing characteristic matching with point cloud data to be detected to obtain an obstacle point cloud;
4) after European clustering processing is carried out on the obstacle point cloud, an obstacle point cloud cluster set is obtained;
5) and finally, outputting the contour boundary and the position of each independent obstacle, thereby realizing the automatic detection of the obstacle on the magnetic suspension track surface.
Preferably, the magnetic suspension track vehicle-mounted mobile scanning system comprises a Z + F laser section scanner, two SICK section scanners, an inertial measurement unit IMU, a GNSS receiver antenna, a GPS clock and a 360-degree panoramic camera;
the Z + F laser section scanners are used for acquiring point cloud data above the magnetic suspension track, and the two SICK section scanners are used for acquiring point cloud data on two sides of the magnetic suspension track; the inertial measurement unit IMU and the matched GNSS receiver antenna are used for providing position and attitude parameters; the GPS clock is used for time synchronization processing; the 360-degree panoramic camera is used for acquiring visual display of the magnetic suspension track image.
Preferably, the step 3) of establishing the K-D tree index specifically includes:
circularly and sequentially taking each dimension (X, Y), (X, Z) and (Y, Z) of the data points, comparing the distribution condition of the data points in each dimension, and if the variance of the coordinate value of a certain dimension is larger, the distribution is more dispersed, and the distribution is more concentrated if the variance is smaller; taking the dimension with large variance as a segmentation dimension, taking the median of the data points in the dimension as a segmentation plane, wherein the plane is vertical to the coordinate axis of the current segmentation dimension, hanging the data points on the left side of the median on the left sub-tree of the median, and hanging the data points on the right side of the median on the right sub-tree of the median; and recursively processing the subtrees until all the data points are mounted, and establishing the K-D tree index for the original reference point cloud data.
Preferably, the step 3) of performing feature matching on the point cloud data to be detected to obtain the obstacle point cloud specifically includes:
the point cloud to be detected gives a query point and a threshold value of a query distance, and a spherical domain is determined by taking the query point as a circle center and the threshold value of the query distance as a radius; performing path sequence or backtracking range search on the left subspace and the right subspace through binary search, and finding out all data with the distance to the query point being less than a threshold value from the original data set; judging whether point cloud data in the original point cloud fall into the ball, if not, indicating that the query point is not in the space range of the original reference point cloud within a set distance threshold, and obtaining the obstacle point cloud; and finally, establishing an index to obtain all the obstacle point cloud sets in the point cloud to be detected.
Preferably, the euclidean clustering of the obstacle point cloud is specifically:
(1) establishing a K-D tree index for the obstacle point cloud;
(2) adding p into a point set cluster from any point p in the obstacle point cloud; searching in the established K-D tree index by taking p as the center and D as the radius, and adding points in the distance range D into a cluster;
(3) repeating the step (2) by taking the points newly added into the cluster as a center p in sequence until no new points are added into the cluster, and filtering the classified point clouds by setting an index tag so as to increase the searching efficiency;
(4) and (3) starting a new clustering, and repeating the steps (1) to (3) from the rest points in the obstacle point cloud until all the points are clustered.
Preferably, the automatic detection of the obstacle on the magnetic suspension track surface is specifically as follows:
and after finishing the obstacle point cloud clustering, filtering the clusters with the number less than a set threshold value, calculating the maximum range of each independent obstacle, and determining the position of each independent obstacle, thereby finishing the detection of the obstacle, and checking the image obtained from the panoramic camera for verifying the correctness of the detection result and confirming the entity type of the obstacle.
Compared with the prior art, the mobile vehicle-mounted laser scanning technology is applied to the detection of the magnetic suspension track obstacles, the defects of large workload and low efficiency of manual detection are overcome, the limitations of a visual sensor and the like are overcome, the detection can be performed in severe environments such as rainy weather and night, the environmental information of the magnetic suspension track can be efficiently acquired, and the magnetic suspension is comprehensively monitored. The detection data obtained each time and the original reference data are compared and analyzed, the running condition of the track can be checked, for example, in obstacle detection, the structural information of the data is fully utilized, the characteristic matching efficiency is accelerated by establishing index search, and finally, the depth information, the position information and the like of the obstacle such as the size, the shape and the like can be obtained by utilizing neighbor domain clustering processing and the like, so that the method has strong environment perception capability and can play an information support role in the operation maintenance safety of magnetic levitation.
Drawings
FIG. 1 is a flow chart of obstacle detection according to the present invention;
FIG. 2 is a schematic diagram of an original reference point cloud;
FIG. 3 is a schematic diagram of a point cloud to be detected;
FIG. 4 is a schematic diagram of the construction and space division of a three-dimensional space K-D;
FIG. 5 is a schematic diagram of Euclidean clustering of obstacle point clouds;
FIG. 6 is a schematic diagram illustrating a detection result of an obstacle in a point cloud to be detected;
fig. 7 is a schematic view of each individual obstacle.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
The invention designs a multi-sensor integrated vehicle-mounted laser scanning system for a magnetic levitation track, and a laser scanner can movably scan and obtain three-dimensional laser point cloud data of a magnetic levitation track surface under a local coordinate system through recording distance, angle, intensity information and the like of a target and integrating with the multi-sensor, so that a magnetic levitation track model can be quickly reconstructed. The method comprises the steps of obtaining original reference point cloud of a magnetic suspension rail surface and point cloud data to be detected through mobile scanning, establishing K-D tree index for the obtained original reference point cloud, performing characteristic matching with the point cloud data to be detected to obtain obstacle point cloud, performing European clustering and other processing on the obstacle point cloud to obtain an obstacle point cloud cluster set, and finally outputting the outline boundary and the position of each independent obstacle, so that automatic detection of the obstacle of the magnetic suspension rail surface is realized, wherein the processing flow of the method is shown in figure 1.
The following introduces the main algorithm and key links involved in the method for detecting the obstacle of the magnetic suspension track surface based on the vehicle-mounted laser scanning system.
1. Design of vehicle-mounted laser scanning system of magnetic levitation track
In order to collect laser point cloud data of a magnetic suspension track, a Z + F laser section scanner, two SICK section scanners, an Inertial Measurement Unit (IMU), a GNSS receiver antenna, a GPS clock, a router, a 360-degree panoramic camera and the like are integrated. The method comprises the steps that point cloud data above and on two sides of a magnetic suspension track are obtained through 3 scanners, position and attitude parameters are provided through combination of an IMU (inertial measurement Unit) and a GNSS (global navigation satellite system) antenna, a GPS (global positioning system) clock is used for time synchronization processing, and a panoramic camera is used for obtaining visual display of images of the magnetic suspension track. After multi-sensor integration and calibration, multi-source data are subjected to time synchronization, space synchronization, point cloud calculation and filtering, and then three-dimensional continuous point cloud of the magnetic suspension track under a local coordinate system can be obtained. The data of the track surface obtained by a Z + F scanner on the vehicle-mounted roof is mainly used.
2. K-D tree index established by original reference point cloud
In order to detect obstacles on a magnetic suspension track surface, firstly, the obstacles need to be defined, namely, objects which are more than the original track environment. Therefore, the scheme obtains two sets of point cloud data, the first set is the original reference point cloud of the magnetic suspension track without the obstacle (as shown in fig. 2), and the second set is the point cloud data to be detected (as shown in fig. 3). In order to perform feature matching on the point cloud to be detected and the original reference point cloud, the data index can be established first by using the structural information of cluster-like clustering morphology and the like presented by the data set, and then the rapid matching is performed, so that the searching efficiency can be greatly accelerated.
The K-D tree (K-D) is a data structure for dividing K-dimensional data space and is mainly applied to searching key data of multi-dimensional space. Aiming at the obtained three-dimensional point cloud data of the original reference of the magnetic suspension track surface without obstacles, firstly, a K-D tree is constructed: circularly and sequentially taking each dimension (X, Y), (X, Z) and (Y, Z) of the data points, comparing the distribution condition of the data points in each dimension, and if the variance of the coordinate value of a certain dimension is larger, the distribution is more dispersed, and the distribution is more concentrated if the variance is smaller; taking the dimension with large variance as a segmentation dimension, taking the median of the data points in the dimension as a segmentation plane, wherein the plane is vertical to the coordinate axis of the current segmentation dimension, hanging the data points on the left side of the median on the left sub-tree of the median, and hanging the data points on the right side of the median on the right sub-tree of the median; and recursively processing the subtrees until all the data points are mounted, and establishing the K-D tree index for the original reference point cloud data, as shown in FIG. 4.
3. Feature matching is carried out on point cloud to be detected
After the index is established for the original reference point cloud, the range of the point cloud to be detected is searched in a K-D tree, and the original point cloud and the point cloud to be detected are subjected to feature matching so as to obtain the obstacle point cloud data. The specific process is as follows: a point cloud to be detected gives a query point and a query distance threshold, and a spherical domain is determined by taking the query point as a circle center and the distance threshold as a radius; performing path sequence or backtracking range search on the left subspace and the right subspace through binary search, and finding out all data with the distance to the query point being less than a threshold value from the original data set; judging whether point cloud data in the original point cloud fall into the ball, if not, indicating that the query point is not in the space range of the original reference point cloud within a certain distance threshold, and obtaining the obstacle point cloud; and finally, establishing an index to obtain all the obstacle point cloud sets in the point cloud to be detected.
4. Cloud and European clustering of obstacle points
After obtaining the obstacle point cloud data, the point cloud needs to be subjected to Euclidean clustering processing to obtain each independent obstacle. The point cloud clustering is to cluster points belonging to the same category into one category according to a certain partition criterion, most of the obtained obstacle point clouds are distributed in a colony, certain distances are reserved between colonies, the point cloud Euclidean clustering used in the invention is to partition different colonies according to the Euclidean distances between point coordinates, and the obstacle point cloud to be partitioned is set as P ═ { P ═ P { (P) }i(Xi,Yi) I |, 1,2,3 … }, and the segmentation distance parameter is D, the specific method is as follows:
(1) establishing K-D tree index for obstacle point cloud
(2) Adding p into a point set cluster from any point p in the obstacle point cloud; taking p as the center and D as the radius, searching in the established K-D tree index, and adding points in the distance range D into the cluster, as shown in figure 5-a;
(3) repeating the step (2) with the point newly added into the cluster as the center p in sequence as shown in figure 5-b until no new point is added into the cluster, and filtering the classified point cloud by setting an index label so as to increase the searching efficiency as shown in figure 5-c;
(4) a new clustering is started and the steps (1) - (3) are repeated from the remaining points in the obstacle point cloud until all the points are clustered, as shown in fig. 5-d.
5. Obstacle detection
After the obstacle point cloud clustering is completed, clustering with too few points is filtered, detection errors caused by instrument scanning, characteristic matching noise and the like are reduced, the maximum range of each independent obstacle is calculated, and the position of each independent obstacle is determined, so that the obstacle detection is completed, images obtained from a panoramic camera are checked, the correctness of a detection result can be verified, and the entity type of the obstacle can be confirmed. As shown in fig. 6, the sizes and distributions of the to-be-detected point cloud relative to the obstacles in the original reference point cloud are divided into two categories, namely person (p) and bird (b), according to the sizes of the detected obstacles, and the detected obstacles are numbered; fig. 7 shows individual obstacle blocks.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (5)
1. A magnetic suspension track obstacle detection method based on point cloud is characterized by comprising the following steps:
1) building a magnetic suspension track vehicle-mounted mobile scanning system;
2) acquiring original reference point cloud of a magnetic suspension rail surface and point cloud data to be detected through mobile scanning;
3) establishing a K-D tree index, and performing characteristic matching with point cloud data to be detected to obtain an obstacle point cloud;
4) after European clustering processing is carried out on the obstacle point cloud, an obstacle point cloud cluster set is obtained;
5) finally, the contour boundary and the position of each independent obstacle are output, and automatic detection of the obstacle on the magnetic suspension track surface is realized;
the step 3) of establishing the K-D tree index specifically comprises the following steps:
circularly and sequentially taking each dimension (X, Y), (X, Z) and (Y, Z) of the data points, comparing the distribution condition of the data points in each dimension, and if the variance of the coordinate value of a certain dimension is larger, the distribution is more dispersed, and the distribution is more concentrated if the variance is smaller; taking the dimension with large variance as a segmentation dimension, taking the median of the data points in the dimension as a segmentation plane, wherein the plane is vertical to the coordinate axis of the current segmentation dimension, hanging the data points on the left side of the median on the left sub-tree of the median, and hanging the data points on the right side of the median on the right sub-tree of the median; and recursively processing the subtrees until all the data points are mounted, and establishing the K-D tree index for the original reference point cloud data.
2. The point cloud-based magnetic levitation track obstacle detection method according to claim 1, wherein the magnetic levitation track vehicle-mounted mobile scanning system comprises a Z + F laser section scanner, two SICK section scanners, an inertial measurement unit IMU, a matched GNSS receiver antenna, a GPS clock and a 360-degree panoramic camera;
the Z + F laser section scanners are used for acquiring point cloud data above the magnetic suspension track, and the two SICK section scanners are used for acquiring point cloud data on two sides of the magnetic suspension track; the inertial measurement unit IMU and the matched GNSS receiver antenna are used for providing position and attitude parameters; the GPS clock is used for time synchronization processing; the 360-degree panoramic camera is used for acquiring visual display of the magnetic suspension track image.
3. The method for detecting obstacles on a magnetic levitation track based on point cloud as claimed in claim 1, wherein the step 3) of obtaining the point cloud of the obstacle by performing feature matching with the point cloud data to be detected specifically comprises:
the point cloud to be detected gives a query point and a threshold value of a query distance, and a spherical domain is determined by taking the query point as a circle center and the threshold value of the query distance as a radius; performing path sequence or backtracking range search on the left subspace and the right subspace through binary search, and finding out all data with the distance to the query point being less than a threshold value from the original data set; judging whether point cloud data in the original point cloud fall into the ball, if not, indicating that the query point is not in the space range of the original reference point cloud within a set distance threshold, and obtaining the obstacle point cloud; and finally, establishing an index to obtain all the obstacle point cloud sets in the point cloud to be detected.
4. The method for detecting the obstacle of the magnetic levitation track based on the point cloud as claimed in claim 1, wherein the Euclidean clustering process for the obstacle point cloud is specifically as follows:
(1) establishing a K-D tree index for the obstacle point cloud;
(2) adding p into a point set cluster from any point p in the obstacle point cloud; searching in the established K-D tree index by taking p as the center and D as the radius, and adding points in the distance range D into a cluster;
(3) repeating the step (2) by taking the points newly added into the cluster as a center p in sequence until no new points are added into the cluster, and filtering the classified point clouds by setting an index tag so as to increase the searching efficiency;
(4) and (3) starting a new clustering, and repeating the steps (1) to (3) from the rest points in the obstacle point cloud until all the points are clustered.
5. A method for detecting obstacles on a magnetic levitation track based on point cloud as claimed in claim 1, wherein the automatic detection of obstacles on the magnetic levitation track surface is specifically:
and after finishing the obstacle point cloud clustering, filtering the clusters with the number less than a set threshold value, calculating the maximum range of each independent obstacle, and determining the position of each independent obstacle, thereby finishing the detection of the obstacle, and checking the image obtained from the panoramic camera for verifying the correctness of the detection result and confirming the entity type of the obstacle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811386577.5A CN109657698B (en) | 2018-11-20 | 2018-11-20 | Magnetic suspension track obstacle detection method based on point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811386577.5A CN109657698B (en) | 2018-11-20 | 2018-11-20 | Magnetic suspension track obstacle detection method based on point cloud |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109657698A CN109657698A (en) | 2019-04-19 |
| CN109657698B true CN109657698B (en) | 2021-09-03 |
Family
ID=66112102
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811386577.5A Active CN109657698B (en) | 2018-11-20 | 2018-11-20 | Magnetic suspension track obstacle detection method based on point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109657698B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110501719B (en) * | 2019-08-27 | 2024-02-09 | 湖南九域同创高分子新材料有限责任公司 | Laser radar-based train obstacle detection method |
| CN110471085B (en) * | 2019-09-04 | 2023-07-04 | 深圳市镭神智能系统有限公司 | Track detecting system |
| CN110889350B (en) * | 2019-11-18 | 2023-05-23 | 四川西南交大铁路发展股份有限公司 | Line obstacle monitoring and alarming system and method based on three-dimensional imaging |
| CN110889871B (en) * | 2019-12-03 | 2021-03-23 | 广东利元亨智能装备股份有限公司 | Robot running method and device and robot |
| CN111178413A (en) * | 2019-12-20 | 2020-05-19 | 浙江欣奕华智能科技有限公司 | 3D point cloud semantic segmentation method, device and system |
| CN111220993B (en) * | 2020-01-14 | 2020-07-28 | 长沙智能驾驶研究院有限公司 | Target scene positioning method and device, computer equipment and storage medium |
| CN111289998A (en) * | 2020-02-05 | 2020-06-16 | 北京汽车集团有限公司 | Obstacle detection method, obstacle detection device, storage medium, and vehicle |
| CN111583257B (en) * | 2020-05-28 | 2023-07-14 | 中国铁道科学研究院集团有限公司 | Railway limit foreign matter intrusion detection method, device and system |
| CN112581521B (en) * | 2020-12-22 | 2023-04-07 | 同济大学 | Method for extracting central line of magnetic suspension track |
| CN112750498B (en) * | 2020-12-30 | 2022-06-24 | 同济大学 | Method for inhibiting HIV virus replication by targeting reverse transcription primer binding site |
| CN112857254B (en) * | 2021-02-02 | 2022-02-08 | 北京大成国测科技有限公司 | Parameter measurement method and device based on unmanned aerial vehicle data and electronic equipment |
| CN113313137A (en) * | 2021-04-01 | 2021-08-27 | 杭州兰亭视觉科技有限责任公司 | Gluing obstacle recognition and removal method |
| CN113406105B (en) * | 2021-06-30 | 2023-04-07 | 湖南凌翔磁浮科技有限责任公司 | Obstacle detection method and system for magnetic levitation detection trolley |
| CN114022760B (en) * | 2021-10-14 | 2022-08-05 | 湖南北斗微芯产业发展有限公司 | Railway tunnel barrier monitoring and early warning method, system, equipment and storage medium |
| CN115082686A (en) * | 2022-08-22 | 2022-09-20 | 同济大学 | Toilet bowl bonding effect detection system and method applying novel three-dimensional point cloud identification method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104931977A (en) * | 2015-06-11 | 2015-09-23 | 同济大学 | Obstacle identification method for smart vehicles |
| CN106406338A (en) * | 2016-04-14 | 2017-02-15 | 中山大学 | Omnidirectional mobile robot autonomous navigation apparatus and method based on laser range finder |
| CN108416257A (en) * | 2018-01-19 | 2018-08-17 | 北京交通大学 | Merge the underground railway track obstacle detection method of vision and laser radar data feature |
| WO2018166747A1 (en) * | 2017-03-15 | 2018-09-20 | Jaguar Land Rover Limited | Improvements in vehicle control |
-
2018
- 2018-11-20 CN CN201811386577.5A patent/CN109657698B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104931977A (en) * | 2015-06-11 | 2015-09-23 | 同济大学 | Obstacle identification method for smart vehicles |
| CN106406338A (en) * | 2016-04-14 | 2017-02-15 | 中山大学 | Omnidirectional mobile robot autonomous navigation apparatus and method based on laser range finder |
| WO2018166747A1 (en) * | 2017-03-15 | 2018-09-20 | Jaguar Land Rover Limited | Improvements in vehicle control |
| CN108416257A (en) * | 2018-01-19 | 2018-08-17 | 北京交通大学 | Merge the underground railway track obstacle detection method of vision and laser radar data feature |
Non-Patent Citations (2)
| Title |
|---|
| 《Online Quadrotor Trajectory Generation and Autonomous Navigation on Point Clouds》;Fei Gao 等;《2016 IEEE International Symposium on Safety,Security,and Rescue Robotics》;20161027;全文 * |
| 《视觉与激光点云融合的深度图像获取方法》;王东敏 等;《军事交通学院学报》;20171031;第19卷(第10期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109657698A (en) | 2019-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109657698B (en) | Magnetic suspension track obstacle detection method based on point cloud | |
| US20240036207A1 (en) | Multiple Resolution, Simultaneous Localization And Mapping Based On 3-D Lidar Measurements | |
| US20220290979A1 (en) | Vehicle localization using the ground surface with an event camera | |
| KR102083909B1 (en) | Automatic extraction method for lane data information for autonomous driving vehicles based on point cloud map | |
| Guan et al. | Automated road information extraction from mobile laser scanning data | |
| Schlichting et al. | Localization using automotive laser scanners and local pattern matching | |
| CN108460416A (en) | A kind of structured road feasible zone extracting method based on three-dimensional laser radar | |
| Steinhauser et al. | Motion segmentation and scene classification from 3D LIDAR data | |
| CN105277190A (en) | Apparatus for a self localization of a vehicle | |
| JP7529835B2 (en) | Self-location estimation device | |
| CN113743171A (en) | Target detection method and device | |
| WO2023060632A1 (en) | Street view ground object multi-dimensional extraction method and system based on point cloud data | |
| CN112285738B (en) | Positioning method and device for rail transit vehicle | |
| CN109444904A (en) | Generation method, device, equipment and the storage medium of assignment test data | |
| Schlichting et al. | Vehicle localization by lidar point correlation improved by change detection | |
| GB2611610A (en) | System and methods for efficient vehicle extent estimation | |
| Kukko et al. | Mobile mapping system and computing methods for modelling of road environment | |
| Gálai et al. | Crossmodal point cloud registration in the Hough space for mobile laser scanning data | |
| Jiang et al. | Effective and robust corrugated beam guardrail detection based on mobile laser scanning data | |
| Nitsch et al. | 3d ground point classification for automotive scenarios | |
| US20230266469A1 (en) | System and method for detecting road intersection on point cloud height map | |
| WO2020118623A1 (en) | Method and system for generating an environment model for positioning | |
| Hofmann et al. | Quality assessment of automatically generated feature maps for future driver assistance systems | |
| US11834065B2 (en) | System, method, and computer program product for detecting road marking points from LiDAR data | |
| Huang et al. | An efficient multi-threshold selection method for lane detection based on lidar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210104 Address after: 200092 Siping Road 1239, Shanghai, Yangpu District Applicant after: TONGJI University Applicant after: CRRC QINGDAO SIFANG Co.,Ltd. Address before: 200092 Siping Road 1239, Shanghai, Yangpu District Applicant before: TONGJI University |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |