CN112613424A - Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium - Google Patents
Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium Download PDFInfo
- Publication number
- CN112613424A CN112613424A CN202011572081.4A CN202011572081A CN112613424A CN 112613424 A CN112613424 A CN 112613424A CN 202011572081 A CN202011572081 A CN 202011572081A CN 112613424 A CN112613424 A CN 112613424A
- Authority
- CN
- China
- Prior art keywords
- obstacle
- cloud data
- point cloud
- determining
- image information
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
Abstract
The embodiment of the invention discloses a rail obstacle detection method, a rail obstacle detection device, electronic equipment and a storage medium. The rail obstacle detection method includes: acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar; determining obstacle image information of the rail area according to the rail area image information; mapping the point cloud data of the rail area to image information of the rail area, and determining the cloud data of the obstacle points according to the image information of the obstacle; determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance. According to the embodiment of the invention, the accuracy of determining the obstacle point cloud data is improved, the detection result of determining the obstacle according to the obstacle point cloud data is more accurate than the detection result directly determined according to the image information, and the accuracy of determining the obstacle detection result in the rail area can be effectively improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of radar, in particular to a rail obstacle detection method and device, electronic equipment and a storage medium.
Background
When the coke quenching car moves on the transportation rail, the obstacle on the rail needs to be detected, and the obstacle is avoided in time, so that the safety of the coke quenching car in the transportation process is ensured.
Most of the existing barrier detection methods depend on a driver to manually check, but certain visual angle blind areas exist in manual checking, and the driver mainly has the responsibility of controlling the vehicle and manually checks to bring threats to the safety of the coke quenching car in the transportation process.
Disclosure of Invention
The embodiment of the invention provides a rail obstacle detection method, a rail obstacle detection device, electronic equipment and a storage medium, so that automation of rail obstacle detection is realized, and the safety of transportation of a coke quenching car on a rail is improved.
In a first aspect, an embodiment of the present invention provides a rail obstacle detection method, including:
acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar;
determining obstacle image information of the rail area according to the rail area image information;
mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle;
determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
In a second aspect, an embodiment of the present invention further provides a rail obstacle detection apparatus, including:
the information acquisition module is used for acquiring rail area image information acquired by the image acquisition device and rail area point cloud data acquired by the laser radar;
the obstacle image determining module is used for determining obstacle image information of the rail area according to the rail area image information;
the obstacle point cloud determining module is used for mapping the point cloud data of the rail area to the image information of the rail area and determining the point cloud data of the obstacle according to the image information of the obstacle;
the obstacle detection module is used for determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
In a third aspect, an embodiment of the present invention further provides an electronic device, including:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a rail obstacle detection method according to any one of the embodiments of the present invention.
In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the rail obstacle detection method according to any one of the embodiments of the present invention.
According to the embodiment of the invention, the point cloud data of the obstacle is determined on the basis of determining the image information of the obstacle according to the image information by combining the image information acquired by the image acquisition device and the point cloud data acquired by the laser radar, so that the accuracy of determining the point cloud data of the obstacle is improved, and the detection result of determining the obstacle according to the point cloud data of the obstacle is more accurate than the detection result directly determined according to the image information.
Drawings
Fig. 1 is a flowchart of a rail obstacle detection method according to a first embodiment of the present invention;
fig. 2 is a flowchart of a rail obstacle detection method according to a second embodiment of the present invention;
fig. 3 is a flowchart of a rail obstacle detection method according to a third embodiment of the present invention;
fig. 4 is a schematic structural view of a rail obstacle detection device according to a fourth embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device in a fifth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a rail obstacle detection method according to a first embodiment of the present invention, which is applicable to detecting an obstacle on a transportation rail of a quenching car. The method may be performed by a rail obstacle detection device, which may be implemented in software and/or hardware and may be configured in an electronic device on the quenching car, for example, the electronic device may be an electronic device with communication and computing capabilities, such as a background server. As shown in fig. 1, the method specifically includes:
The image acquisition device is used for acquiring image information on the rail, and the image acquisition device can be a monitoring camera and the like. For example, the image acquisition device is arranged on a quenching car, and the image information of the rail of the quenching car in the moving process is acquired. The laser radar is used for acquiring three-dimensional point cloud data of a surrounding area, for example, the laser radar and the image acquisition device are arranged on a quenching car together to acquire the three-dimensional point cloud data of a rail area.
Specifically, monitoring cameras and laser radars are arranged at the car head parts on the two sides of the coke quenching car, and when the coke quenching car moves on a rail, image information and three-dimensional point cloud data of the rail area of the coke quenching car in the moving process are obtained simultaneously. The obstacles in the rail area can be identified through the image information and the three-dimensional point cloud data.
And 102, determining obstacle image information of the rail area according to the rail area image information.
Wherein, the obstacle is an object which is positioned in the rail area and causes interference to the movement of the quenching car. For example, the obstacle may be a pedestrian, another quench car traveling on the same rail, another vehicle, or a barricade, etc.
Specifically, the image information of the obstacle located in the rail area in the image is identified by identifying the image captured by the monitoring camera. For example, the boundary information of the obstacle in the image is obtained by an image recognition technique, for example, the image is input into a pre-trained obstacle recognition model, and the result of the obstacle information in the image recognized by the model is obtained. Obstacle image information is information of the position of an obstacle on an image.
And unifying the coordinates of the point cloud data of the rail area and the coordinates of the image information of the rail area to map the point cloud data into the image information so that the point cloud data and the image information are matched. And determining point cloud data mapped with the image information according to the image information of the obstacle as the cloud data of the obstacle points. Illustratively, when the obstacle image information is boundary information of an obstacle on the image, the obstacle point cloud data is all point cloud data determined according to the mapping result, including the point cloud data on the boundary.
The method has the advantages that the obstacle point cloud data after mapping is determined based on the obstacle image information, the accuracy of point cloud data determination is guaranteed, the efficiency of point cloud data determination is improved by determining the point cloud data according to the image information due to the fact that the image information determination is efficient and the steps are simple, and the complex steps that the point cloud data of the rail area needs to be integrally processed are avoided.
Since each point in the obstacle point cloud data includes a three-dimensional coordinate, the height, width, and distance of the obstacle from the lidar sensor, i.e., the distance of the obstacle to the quench car, may be determined from the obstacle point cloud data.
Illustratively, three-dimensional reconstruction is carried out through the obstacle point cloud data to obtain a three-dimensional model of the obstacle, and specific size information and distance information of the obstacle can be obtained through the three-dimensional model.
In an optional embodiment, the method further comprises acquiring position information acquired by the millimeter wave radar;
correspondingly, mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle, wherein the method comprises the following steps:
mapping the point cloud data and the position information of the rail area to image information of the rail area, and determining the consistency of mapping results;
if the image information of the obstacle is consistent with the image information of the obstacle, determining obstacle point cloud data according to the obstacle image information;
otherwise, determining the position information of the obstacle according to the position information acquired by the millimeter wave radar.
Under the condition of bad weather such as heavy fog, heavy rain and the like, certain errors and errors exist in point cloud data acquired through an image acquisition device or a laser radar. Therefore, in the embodiment of the invention, the position information acquired by the millimeter wave radar is acquired, and the millimeter wave radar can be arranged on the coke quenching car for example and acquires the position information of the same rail area as the laser radar. The millimeter wave radar acquires two-dimensional coordinate information of an obstacle in the rail area, the two-dimensional coordinate information and the three-dimensional coordinate information acquired by the laser radar are mapped into the rail area image information, if the mapping results are consistent, the obstacle is detected by the laser radar and the millimeter wave radar, and therefore obstacle point cloud data are determined according to the obstacle image information. The obstacle detection method has the advantages that detection is carried out in multiple modes, and the obstacle determination accuracy is improved.
If the mapping results are inconsistent, omission or errors exist in the point cloud data, and therefore obstacle position information is determined according to the position information acquired by the millimeter wave radar so as to guarantee the accuracy of the obstacle detection result. For example, in an optional embodiment, if the image information of the rail area collected by the image detection device does not detect the image information of the obstacle or the point cloud data of the rail area collected by the laser radar is smaller than a certain threshold, the position information of the obstacle is determined according to the position information collected by the millimeter wave radar. So as to ensure the suitability of rail obstacle detection in various weathers.
In the embodiment of the invention, the obstacle is detected by combining the information acquired by the image acquisition device, the laser radar and the millimeter wave radar, so that the accuracy of the detection result is improved.
According to the embodiment of the invention, the point cloud data of the obstacle is determined on the basis of determining the image information of the obstacle according to the image information by combining the image information acquired by the image acquisition device and the point cloud data acquired by the laser radar, so that the accuracy of determining the point cloud data of the obstacle is improved, and the detection result of determining the obstacle according to the point cloud data of the obstacle is more accurate than the detection result directly determined according to the image information.
Example two
Fig. 2 is a flowchart of a rail obstacle detection method according to a second embodiment of the present invention, which is further optimized based on the first embodiment, where the obstacles include pedestrian obstacles and other obstacles. As shown in fig. 2, the method includes:
And step 202, determining image information of the pedestrian obstacles in the rail area according to the image information of the rail area based on a pre-trained deep learning network model.
Because a huge training data set exists for the pedestrian barrier, a pre-trained deep learning network model can be obtained based on the published pedestrian data set, and the accuracy of the pedestrian detection result of the model can be ensured.
Specifically, training is carried out by adopting yolov3 network according to the published data set to obtain a deep learning network model, the image information of the rail area is used as the input of the deep learning network model, and the pedestrian prediction result in the image is output, thereby determining the image information of the pedestrian obstacles in the rail area.
And unifying the coordinates of the point cloud data of the rail area and the coordinates of the image information of the rail area to map the point cloud data into the image information so that the point cloud data and the image information are matched. And determining point cloud data mapped with the pedestrian image information according to the pedestrian obstacle image information to serve as pedestrian obstacle point cloud data. Illustratively, point cloud data at the position is determined as pedestrian obstacle point cloud data according to boundary position information of a pedestrian output by the deep learning network model.
In an alternative embodiment, the pedestrian obstacle image information is pedestrian obstacle boundary information;
mapping the point cloud data of the rail area to image information of the rail area, and determining the point cloud data in the boundary of the pedestrian obstacle according to the boundary information of the pedestrian obstacle;
clustering point cloud data in the pedestrian barrier boundary to obtain a clustering result;
if the clustering result is one type, determining point cloud data in the boundary of the pedestrian obstacle as pedestrian obstacle point cloud data;
otherwise, determining the pedestrian obstacle point cloud data according to the distance of each type of point cloud data in the clustering result and/or the characteristics of the pedestrian obstacle.
The pedestrian obstacle boundary information refers to coordinate position information in a pedestrian prediction result in an image output by a model, and the boundary of the pedestrian obstacle in the image can be determined according to the coordinate position information.
And mapping the point cloud data of the rail area to the image information of the rail area, and determining the point cloud data in the boundary of the pedestrian obstacle according to the boundary information of the pedestrian obstacle. Since the point cloud data includes three-dimensional information, when there is another object in front of or behind the pedestrian obstacle in the image, the point cloud data of the object is also included in the point cloud data in the boundary of the pedestrian obstacle. If all the point clouds determined according to the boundary information are directly used as the pedestrian obstacle point clouds, the pedestrian obstacle point clouds are determined inaccurately.
And clustering the point cloud data in the boundary of the pedestrian obstacle, wherein the clustering operation can realize the function of classifying the point clouds in the boundary of the pedestrian obstacle according to information such as the distance between the point clouds in the point cloud data, and distinguish objects represented by the point clouds in the region.
When the clustering result is a type, the type is the pedestrian because the clustering result is the point cloud determined by the pedestrian image obtained by predicting according to the deep learning network model, the type does not contain the point cloud data of other objects in the region, and the point cloud data in the boundary of the pedestrian obstacle is determined as the pedestrian obstacle point cloud data.
Otherwise, when the clustering result exceeds one type, the object represented by the point cloud data in the boundary of the pedestrian obstacle is not only a pedestrian, and then the pedestrian obstacle point cloud data is determined according to the distance of each type of point cloud data and/or the characteristics of the pedestrian obstacle. Exemplarily, the distance between each type of object and the laser radar is determined according to the clustering result, the object with the closest distance can be used as a pedestrian, and the point cloud with the closest distance is the pedestrian obstacle point cloud; or identifying each class object according to the height and width information of the pedestrian and the outline shape of the pedestrian, judging to obtain a clustering result meeting the pedestrian standard as the pedestrian, and using the class point cloud as the pedestrian obstacle point cloud.
The accuracy of the pedestrian obstacle point cloud determination result can be further ensured by analyzing according to the clustering result, and the point clouds of other obstacles in the determined pedestrian obstacle point cloud are prevented from influencing the determination of the subsequent detection result.
And step 204, determining other obstacle point cloud data according to the rail area point cloud data and the pedestrian obstacle point cloud data.
Specifically, the pedestrian obstacle point cloud data in the point cloud data of the rail area is filtered, and the remaining point cloud data is the other obstacle point cloud data.
And step 205, clustering point cloud data of other obstacles to obtain clustering results of other obstacles.
Because other obstacles comprise various types, the characteristics are not clear, and if the same method as the method for determining the pedestrian obstacle point cloud is adopted, namely image information of other obstacles is determined by using an image recognition technology, and then point cloud data is determined based on the image information, the fact that the determination of the image information of other obstacles is inaccurate can be caused, and the determination of subsequent point cloud data is directly influenced. And because the types and the characteristics of other obstacles are not clear, a large number of training sample sets cannot be acquired or the acquisition difficulty is high, so that the prediction result of the deep learning neural network is inaccurate and the efficiency is low.
Therefore, the cluster analysis is directly carried out on the other obstacle point cloud data with the pedestrian obstacle point cloud data filtered out, each class in the obtained cluster result represents one other obstacle, and therefore the point cloud data of each other obstacle can be obtained through clustering. The accuracy of determining the cloud data of each other obstacle point can be improved by clustering all the cloud data of other obstacle points.
And step 206, determining point cloud data and classification results of other obstacles according to the clustering results of other obstacles and the characteristics of various obstacles.
And (4) classifying and judging each other obstacle in the clustering results of other obstacles according to the overall characteristics of each obstacle to obtain a classification result, wherein the point cloud data of each other obstacle is the point cloud data of each other obstacle. Illustratively, on the basis of the above example, on the rail transported by the quenching car, the other obstacles at least comprise other quenching cars, collision avoidance barriers, roadblocks and the like, and the classification results of the other obstacles and the corresponding point cloud data can be obtained by analyzing the general shape of each type of obstacle.
After the point cloud data of each obstacle in the rail area is determined, the detection result of each obstacle can be obtained according to the point cloud data of each obstacle. Specifically, the height, width and distance of the pedestrian obstacle can be determined according to the pedestrian obstacle point cloud data, and the height, width and distance of each other obstacle can be obtained according to the point cloud data of each other obstacle in the other obstacle point cloud data.
According to the embodiment of the invention, the pedestrian obstacle is detected through the pre-trained deep learning network model, so that the accuracy of cloud data of the pedestrian obstacle point is ensured, and the accuracy of determination of cloud data of other obstacle points is further ensured. And point cloud data of other obstacles in the point cloud data of other obstacles is obtained based on clustering analysis, so that the detection result of the obstacle is determined according to the point cloud data of the pedestrian obstacle and the point cloud data of the other obstacles, and the accuracy and the efficiency of determining the detection result are improved.
EXAMPLE III
Fig. 3 is a flowchart of a rail obstacle detection method according to a third embodiment of the present invention, which is further optimized based on the second embodiment. As shown in fig. 3, the method includes:
301, acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar.
The higher the resolution of the laser radar is, the larger the acquisition range is, the denser the point cloud is, the point cloud data of the rail area directly acquired by the laser radar can comprise irrelevant point clouds of a non-rail area, and the more the point clouds of the rail area are, the processing efficiency of the point cloud data can be influenced, so that the point cloud data of the rail area directly acquired by the laser radar is preprocessed before the point cloud data are processed, and the efficiency of subsequent operations such as point cloud mapping and clustering is improved. It can be seen that the preprocessing steps such as step 303 may be performed before the processing of the point cloud data of the rail area, that is, after step 302, or in synchronization with step 302.
The voxelization filtering is to perform down-sampling on the point cloud data, that is, to reduce the number of the point clouds and maintain the shape characteristics of the point clouds, and to improve the point cloud processing efficiency by reducing the density of the unit distance of the point clouds. Illustratively, a three-dimensional voxel grid (which can be thought of as a collection of tiny spatial three-dimensional cubes) is created from the input point cloud data of the rail region, and then, in each voxel (i.e., three-dimensional cube), the center of gravity of all points in the voxel is used to approximately display other points in the voxel, so that all points in the voxel are finally represented by a center of gravity point, thereby obtaining the filtered point cloud data.
Prior to step 303, filtering the rail area point cloud data in the non-detection space is also included.
As can be seen from the above, the point cloud data of the rail area directly acquired by the laser radar may include unrelated point clouds of the non-rail area, such as an obstacle, ground debris, and a noise point outside the safe distance, and therefore, the unrelated point clouds of the non-rail area need to be filtered out.
The detection space is used for controlling the detection range, the detection space of the rail area is determined according to the preset upper boundary threshold, the preset left boundary threshold and the preset right boundary threshold, the point cloud in the non-detection space is removed, the efficiency of point cloud processing can be improved by filtering the point cloud data of the rail area in the non-detection space, and the problem that the point cloud processing is difficult due to the existence of excessive irrelevant point clouds is avoided.
And step 304, determining a normal vector of the filtered point cloud data.
And solving the normal vector of each point cloud in the filtered point cloud data. The normal vector of the point cloud is approximate to a tangent plane normal of the estimation surface, the point cloud in the neighborhood is required to be determined, and the size of the neighborhood is generally represented by the neighborhood radius value or the number of the adjacent points. An excessively large neighborhood can smear the details of the three-dimensional structure, so that the normal vector is excessively rough, and an excessively small neighborhood contains too few points and is strongly interfered by noise. In the embodiment of the present invention, values need to be obtained according to factors such as the dot resolution, the degree of detail of the object, and the application, and the like, and the present invention is not limited herein.
And 305, determining ground point cloud data in the filtered point cloud data according to the normal vector, and determining fitted ground according to the ground point cloud data.
And determining the range and the direction of the ground point cloud according to a pre-calibrated result, determining the point cloud with the normal vector meeting the conditions as ground point cloud data according to the range and the direction, and fitting a plane as a fitting ground according to the ground point cloud data.
For example, a determination instruction of the user for the ground point cloud range may also be received, the range of the ground point cloud is determined according to the determination instruction, and the ground point cloud data is determined according to the normal vector in the range. The normal vector of the ground points to the Z axis of the coordinate axis calibrated in advance, and the specific value range can be determined according to the coordinate system of the laser radar.
The left track and the right track are arranged at a certain height from the ground, and the tracks can be regarded as a straight line within a certain distance, so that a normal vector of a point cloud of the left track and the right track is determined, for example, the normal vector can be set to be the same as the normal vector of the point cloud of the ground, the height of the left track and the right track from the fitting ground is determined, the point cloud meeting the conditions is determined to be rail track point cloud data according to the normal vector and the height, rail left track point cloud data and rail right track point cloud data can be determined according to the positions of the point clouds, and a fitting left track plane and a fitting right track plane which are perpendicular to the fitting ground are determined according to the.
And 307, filtering point cloud data of which the normal vector is out of a preset normal vector range, and determining the filtered point cloud data so that the normal vector of the filtered point cloud data points to the image acquisition device and the laser radar.
Since one object comprises a plurality of surfaces, for example, for other quenching cars in an obstacle, the front, the left, the right and even the upper point cloud data can be obtained, and the height, the width and the distance of the quenching car can be estimated according to the point cloud data on the front, the point cloud data on the other surfaces can cause low point cloud processing efficiency and bring certain difficulty.
A normal vector range is predetermined, and the point cloud in the normal vector range belongs to the front point cloud of the obstacle. In the embodiment of the invention, the surface of the obstacle facing the laser radar and the image acquisition device is used as the front surface. And filtering out the point clouds of which the normal vectors are not in the preset normal vector range so as to reduce the number of the point clouds and improve the point cloud processing efficiency.
And 308, projecting the filtered point cloud data to the fitting ground to obtain a projection result of each point cloud data.
And projecting the processed point cloud onto the fitting ground to obtain point cloud data projected on the ground.
309, filtering point cloud data of which the projection result is outside a preset area range, and determining the projected point cloud data; wherein the area range is greater than the range determined from the fitted left orbital plane and the fitted right orbital plane.
In the embodiment of the present invention, the area range is larger than the range determined according to the fitting left track plane and the fitting right track plane, for example, the left boundary of the area range is at a preset distance on the left side of the fitting left track plane, the right boundary is at a preset distance on the right side of the fitting right track plane, the preset distance may be determined according to the size of an obstacle appearing in the rail area in an actual scene, and no limitation is made here.
And filtering the point cloud of which the ground projection result is outside the preset area range, and processing the point cloud of which the ground projection result is inside the preset area range, so that the obtained obstacles can be ensured to be in the rail area and can be obstacles which can cause certain influence on transportation of the coke quenching car on the rail.
The operation from step 303 to step 309 can be performed as the operation of preprocessing the point cloud data acquired by the laser radar, and the point cloud data of the irrelevant obstacles in the rail area can be effectively filtered out by preprocessing the point cloud data, so that the efficiency of determining the obstacle detection result in the rail area is improved.
And 310, mapping the projected point cloud data to the rail area image information, and determining obstacle point cloud data according to the obstacle image information.
And mapping the projected point cloud data to the rail area image information, and determining the point cloud data of each obstacle according to the obstacle image information.
The point cloud data of each obstacle and the position information of the fitted left orbit plane and the fitted right orbit plane are determined, and exemplarily, whether the point cloud data of each obstacle is located inside or outside the orbit and which orbit is located outside is determined. And if the point cloud data part of any obstacle is positioned in the track, the position of the point cloud data of the obstacle is in the track.
And if the position of the obstacle point cloud data is outside the left track, determining the height, width and distance of the obstacle according to the fitted ground and the obstacle point cloud data, and determining the distance from the obstacle to the left track according to the fitted left track plane.
And when the obstacle point cloud data are positioned outside the left track, analyzing the obstacle point cloud data to obtain the height, width and distance of an obstacle represented by the obstacle point cloud data and the distance from a plane of the fitted left track, wherein the distance of the obstacle is the distance from the laser radar. Illustratively, three-dimensional reconstruction is carried out according to the obstacle point cloud data to obtain a three-dimensional model of the obstacle, and the height, the width and the distance of the obstacle and the distance of the distance fitting left orbit plane can be obtained according to the three-dimensional model.
And if the position of the obstacle point cloud data is outside the right track, determining the height, width and distance of the obstacle according to the fitted ground and the obstacle point cloud data, and determining the distance from the obstacle to the right track according to the fitted right track plane.
And when the obstacle point cloud data are positioned outside the right track, analyzing the obstacle point cloud data to obtain the height, width and distance of an obstacle represented by the obstacle point cloud data and the distance from a plane of the right track, wherein the distance of the obstacle is the distance from the laser radar. Illustratively, three-dimensional reconstruction is carried out according to the obstacle point cloud data to obtain a three-dimensional model of the obstacle, and the height, the width and the distance of the obstacle and the distance of the distance fitting right orbit plane can be obtained according to the three-dimensional model.
And if the position of the obstacle point cloud data is in the track, determining the height, width and distance of the obstacle according to the fitted ground and the obstacle point cloud data.
And when the obstacle point cloud data are positioned in the track, analyzing the obstacle point cloud data to obtain the height, width and distance of an obstacle represented by the obstacle point cloud data, wherein the distance of the obstacle is the distance from the laser radar. Illustratively, three-dimensional reconstruction is carried out according to the obstacle point cloud data to obtain a three-dimensional model of the obstacle, and the height, the width and the distance of the obstacle can be obtained according to the three-dimensional model.
Different detection results are obtained according to different positions of the obstacle point cloud, for example, when the obstacle point cloud is positioned in a track, the normal transportation of the coke quenching car on a rail is influenced certainly, and the distance between the coke quenching car and the track is not required to be determined; when the obstacle is positioned outside the track, the size of the obstacle and the distance from the obstacle to the track can be determined, whether the obstacle influences the transportation of the coke quenching car or not is determined according to the distance and the size, and the determination efficiency of the detection result can be effectively improved.
According to the embodiment of the invention, the cloud data of the irrelevant obstacle points in the rail area can be effectively filtered by preprocessing the point cloud data, so that the efficiency of determining the obstacle detection result in the rail area is improved. The point cloud data after pretreatment is processed, so that the processing efficiency of the point cloud data is further improved.
Example four
Fig. 4 is a schematic structural diagram of a rail obstacle detection device in a fourth embodiment of the invention, which is applicable to detecting an obstacle on a transportation rail of a coke quenching car. As shown in fig. 4, the apparatus includes:
the information acquisition module 410 is used for acquiring rail area image information acquired by the image acquisition device and rail area point cloud data acquired by the laser radar;
the obstacle image determining module 420 is configured to determine obstacle image information of the rail area according to the rail area image information;
an obstacle point cloud determining module 430, configured to map the rail area point cloud data into the rail area image information, and determine the obstacle point cloud data according to the obstacle image information;
an obstacle detection module 440, configured to determine a detection result of an obstacle according to the obstacle point cloud data, where the detection result includes at least one of: height, width and distance.
According to the embodiment of the invention, the point cloud data of the obstacle is determined on the basis of determining the image information of the obstacle according to the image information by combining the image information acquired by the image acquisition device and the point cloud data acquired by the laser radar, so that the accuracy of determining the point cloud data of the obstacle is improved, and the detection result of determining the obstacle according to the point cloud data of the obstacle is more accurate than the detection result directly determined according to the image information.
Optionally, the obstacle comprises a pedestrian;
the obstacle image determination module 420 is specifically configured to:
determining pedestrian obstacle image information of the rail area according to the rail area image information based on a pre-trained deep learning network model;
accordingly, the obstacle point cloud determination module 430 includes:
and the pedestrian obstacle point cloud determining unit is used for mapping the point cloud data of the rail area to the image information of the rail area and determining the cloud data of the pedestrian obstacle point according to the image information of the pedestrian obstacle.
Optionally, the image information of the pedestrian obstacle is boundary information of the pedestrian obstacle;
pedestrian obstacle point cloud determination unit, specifically used for:
mapping the point cloud data of the rail area to the image information of the rail area, and determining the point cloud data in the boundary of the pedestrian obstacle according to the boundary information of the pedestrian obstacle;
clustering the point cloud data in the pedestrian obstacle boundary to obtain a clustering result;
if the clustering result is of one type, determining the point cloud data in the boundary of the pedestrian obstacle as the cloud data of the pedestrian obstacle point;
otherwise, determining the pedestrian obstacle point cloud data according to the distance of each type of point cloud data in the clustering result and/or the characteristics of the pedestrian obstacle.
Optionally, the obstacle further comprises other obstacles;
correspondingly, the device further comprises an other obstacle point cloud determination module for:
determining other obstacle point cloud data according to the rail area point cloud data and the pedestrian obstacle point cloud data;
clustering the point cloud data of other obstacles to obtain clustering results of other obstacles;
and determining point cloud data and classification results of other obstacles according to the clustering results of the other obstacles and the characteristics of various obstacles.
Optionally, the apparatus further includes a point cloud data preprocessing module, configured to:
performing voxelization filtering on the point cloud data of the rail area to obtain filtered point cloud data;
determining a normal vector of the filtered point cloud data;
determining ground point cloud data in the filtered point cloud data according to the normal vector, and determining fitted ground according to the ground point cloud data;
determining the point cloud data of the left rail track and the point cloud data of the right rail track in the filtered point cloud data according to the normal vector and the fitting ground, and determining a fitting left rail plane and a fitting right rail plane according to the point cloud data of the left rail track and the point cloud data of the right rail track; wherein the fitted left orbit plane and the fitted right orbit plane are perpendicular to the fitted ground;
filtering point cloud data of which the normal vector is outside a preset normal vector range, and determining the filtered point cloud data so that the normal vector of the filtered point cloud data points to the image acquisition device and the laser radar;
projecting the filtered point cloud data to the fitting ground to obtain a projection result of each point cloud data;
filtering the point cloud data of the projection result outside a preset area range, and determining the projected point cloud data; wherein the region range is greater than a range determined from the fitted left orbital plane and the fitted right orbital plane;
correspondingly, the obstacle point cloud determining module 430 is specifically configured to:
and mapping the projected point cloud data to the rail area image information, and determining the obstacle point cloud data according to the obstacle image information.
Optionally, the apparatus further comprises an obstacle position determination module, configured to:
determining the position of the obstacle point cloud data according to the fitted left orbit plane and the fitted right orbit plane; wherein the location is included within the track, outside the left track, and outside the right track;
if the position of the obstacle point cloud data is outside the left rail, determining the height, width and distance of an obstacle according to the fitting ground and the obstacle point cloud data, and determining the distance between the obstacle and the left rail according to the fitting left rail plane;
if the position of the obstacle point cloud data is outside the right track, determining the height, width and distance of an obstacle according to the fitting ground and the obstacle point cloud data, and determining the distance between the obstacle and the right track according to the fitting right track plane;
and if the position of the obstacle point cloud data is in the track, determining the height, width and distance of the obstacle according to the fitted ground and the obstacle point cloud data.
Optionally, the apparatus further includes a millimeter wave radar module, configured to obtain position information acquired by the millimeter wave radar;
correspondingly, the obstacle point cloud determination module is specifically configured to:
mapping the point cloud data of the rail area and the position information to the image information of the rail area, and determining the consistency of mapping results;
if the image information of the obstacle is consistent with the image information of the obstacle, determining the cloud data of the obstacle point according to the image information of the obstacle;
otherwise, determining the position information of the obstacle according to the position information acquired by the millimeter wave radar.
The rail obstacle detection device provided by the embodiment of the invention can execute the rail obstacle detection method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects for executing the rail obstacle detection method.
EXAMPLE five
Fig. 5 is a schematic structural diagram of an electronic device according to a fifth embodiment of the present invention. FIG. 5 illustrates a block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 5 is only an example and should not bring any limitation to the function and the scope of use of the embodiment of the present invention.
As shown in FIG. 5, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory device 28, and a bus 18 that couples various system components including the system memory device 28 and the processing unit 16.
The system storage 28 may include computer system readable media in the form of volatile storage, such as Random Access Memory (RAM) 30 and/or cache storage 32. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 5, and commonly referred to as a "hard drive"). Although not shown in FIG. 5, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Storage 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.
A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in storage 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.
The processing unit 16 executes various functional applications and data processing by running a program stored in the system storage device 28, for example, to implement a rail obstacle detection method provided by an embodiment of the present invention, including:
acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar;
determining obstacle image information of the rail area according to the rail area image information;
mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle;
determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
EXAMPLE six
An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a rail obstacle detection method according to an embodiment of the present invention, where the computer program includes:
acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar;
determining obstacle image information of the rail area according to the rail area image information;
mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle;
determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, or the like, as well as conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. A rail obstacle detection method, comprising:
acquiring rail area image information acquired by an image acquisition device and rail area point cloud data acquired by a laser radar;
determining obstacle image information of the rail area according to the rail area image information;
mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle;
determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
2. The method of claim 1, wherein the obstacle comprises a pedestrian;
determining obstacle image information for the rail area from the rail area image information, comprising:
determining pedestrian obstacle image information of the rail area according to the rail area image information based on a pre-trained deep learning network model;
correspondingly, mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle, including:
and mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the pedestrian obstacle points according to the image information of the pedestrian obstacle.
3. The method according to claim 2, wherein the pedestrian obstacle image information is pedestrian obstacle boundary information;
mapping the rail area point cloud data to the rail area image information, and determining the pedestrian obstacle point cloud data according to the pedestrian obstacle image information, wherein the method comprises the following steps:
mapping the point cloud data of the rail area to the image information of the rail area, and determining the point cloud data in the boundary of the pedestrian obstacle according to the boundary information of the pedestrian obstacle;
clustering the point cloud data in the pedestrian obstacle boundary to obtain a clustering result;
if the clustering result is of one type, determining the point cloud data in the boundary of the pedestrian obstacle as the cloud data of the pedestrian obstacle point;
otherwise, determining the pedestrian obstacle point cloud data according to the distance of each type of point cloud data in the clustering result and/or the characteristics of the pedestrian obstacle.
4. The method of claim 2, wherein the obstacles further comprise other obstacles;
correspondingly, in mapping the point cloud data of the rail area to the image information of the rail area, after determining the point cloud data of the pedestrian obstacle according to the image information of the pedestrian obstacle, the method further comprises:
determining other obstacle point cloud data according to the rail area point cloud data and the pedestrian obstacle point cloud data;
clustering the point cloud data of other obstacles to obtain clustering results of other obstacles;
and determining point cloud data and classification results of other obstacles according to the clustering results of the other obstacles and the characteristics of various obstacles.
5. The method of claim 1, further comprising, prior to determining the obstacle point cloud data from the obstacle image information in mapping the rail area point cloud data to the rail area image information:
performing voxelization filtering on the point cloud data of the rail area to obtain filtered point cloud data;
determining a normal vector of the filtered point cloud data;
determining ground point cloud data in the filtered point cloud data according to the normal vector, and determining fitted ground according to the ground point cloud data;
determining the point cloud data of the left rail track and the point cloud data of the right rail track in the filtered point cloud data according to the normal vector and the fitting ground, and determining a fitting left rail plane and a fitting right rail plane according to the point cloud data of the left rail track and the point cloud data of the right rail track; wherein the fitted left orbit plane and the fitted right orbit plane are perpendicular to the fitted ground;
filtering point cloud data of which the normal vector is outside a preset normal vector range, and determining the filtered point cloud data so that the normal vector of the filtered point cloud data points to the image acquisition device and the laser radar;
projecting the filtered point cloud data to the fitting ground to obtain a projection result of each point cloud data;
filtering the point cloud data of the projection result outside a preset area range, and determining the projected point cloud data; wherein the region range is greater than a range determined from the fitted left orbital plane and the fitted right orbital plane;
correspondingly, mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle, including:
and mapping the projected point cloud data to the rail area image information, and determining the obstacle point cloud data according to the obstacle image information.
6. The method of claim 5, further comprising, after mapping the projected point cloud data to the rail area image information, determining the obstacle point cloud data from the obstacle image information:
determining the position of the obstacle point cloud data according to the fitted left orbit plane and the fitted right orbit plane; wherein the location is included within the track, outside the left track, and outside the right track;
if the position of the obstacle point cloud data is outside the left rail, determining the height, width and distance of an obstacle according to the fitting ground and the obstacle point cloud data, and determining the distance between the obstacle and the left rail according to the fitting left rail plane;
if the position of the obstacle point cloud data is outside the right track, determining the height, width and distance of an obstacle according to the fitting ground and the obstacle point cloud data, and determining the distance between the obstacle and the right track according to the fitting right track plane;
and if the position of the obstacle point cloud data is in the track, determining the height, width and distance of the obstacle according to the fitted ground and the obstacle point cloud data.
7. The method according to claim 1, further comprising obtaining position information collected by the millimeter wave radar;
correspondingly, mapping the point cloud data of the rail area to the image information of the rail area, and determining the cloud data of the obstacle point according to the image information of the obstacle, including:
mapping the point cloud data of the rail area and the position information to the image information of the rail area, and determining the consistency of mapping results;
if the image information of the obstacle is consistent with the image information of the obstacle, determining the cloud data of the obstacle point according to the image information of the obstacle;
otherwise, determining the position information of the obstacle according to the position information acquired by the millimeter wave radar.
8. A rail obstacle detection device, comprising:
the information acquisition module is used for acquiring rail area image information acquired by the image acquisition device and rail area point cloud data acquired by the laser radar;
the obstacle image determining module is used for determining obstacle image information of the rail area according to the rail area image information;
the obstacle point cloud determining module is used for mapping the point cloud data of the rail area to the image information of the rail area and determining the point cloud data of the obstacle according to the image information of the obstacle;
the obstacle detection module is used for determining a detection result of the obstacle according to the obstacle point cloud data, wherein the detection result comprises at least one of the following items: height, width and distance.
9. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a rail obstacle detection method according to any one of claims 1-7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a rail obstacle detection method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011572081.4A CN112613424A (en) | 2020-12-27 | 2020-12-27 | Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011572081.4A CN112613424A (en) | 2020-12-27 | 2020-12-27 | Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112613424A true CN112613424A (en) | 2021-04-06 |
Family
ID=75249286
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011572081.4A Pending CN112613424A (en) | 2020-12-27 | 2020-12-27 | Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112613424A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113570622A (en) * | 2021-07-26 | 2021-10-29 | 北京全路通信信号研究设计院集团有限公司 | Obstacle determination method and device, electronic equipment and storage medium |
| CN113807795A (en) * | 2021-10-19 | 2021-12-17 | 上海擎朗智能科技有限公司 | Method for identifying congestion in robot distribution scene, robot and distribution system |
| CN113920487A (en) * | 2021-10-09 | 2022-01-11 | 广东汇天航空航天科技有限公司 | Obstacle detection processing method, device and system |
| CN115598656A (en) * | 2022-12-14 | 2023-01-13 | 成都运达科技股份有限公司（Cn） | Obstacle detection method, device and system based on suspension track |
| CN115607052A (en) * | 2022-12-19 | 2023-01-17 | 科大讯飞股份有限公司 | Cleaning method, device and equipment of robot and cleaning robot |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109116374A (en) * | 2017-06-23 | 2019-01-01 | 百度在线网络技术（北京）有限公司 | Determine the method, apparatus, equipment and storage medium of obstacle distance |
| CN110096059A (en) * | 2019-04-25 | 2019-08-06 | 杭州飞步科技有限公司 | Automatic Pilot method, apparatus, equipment and storage medium |
| CN111160302A (en) * | 2019-12-31 | 2020-05-15 | 深圳一清创新科技有限公司 | Obstacle information identification method and device based on automatic driving environment |
| CN112101092A (en) * | 2020-07-31 | 2020-12-18 | 北京智行者科技有限公司 | Automatic driving environment sensing method and system |
-
2020
- 2020-12-27 CN CN202011572081.4A patent/CN112613424A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109116374A (en) * | 2017-06-23 | 2019-01-01 | 百度在线网络技术（北京）有限公司 | Determine the method, apparatus, equipment and storage medium of obstacle distance |
| CN110096059A (en) * | 2019-04-25 | 2019-08-06 | 杭州飞步科技有限公司 | Automatic Pilot method, apparatus, equipment and storage medium |
| CN111160302A (en) * | 2019-12-31 | 2020-05-15 | 深圳一清创新科技有限公司 | Obstacle information identification method and device based on automatic driving environment |
| CN112101092A (en) * | 2020-07-31 | 2020-12-18 | 北京智行者科技有限公司 | Automatic driving environment sensing method and system |
Non-Patent Citations (1)
| Title |
|---|
| 张耀威等: "基于图像与点云的三维障碍物检测", 《计算机工程与设计》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113570622A (en) * | 2021-07-26 | 2021-10-29 | 北京全路通信信号研究设计院集团有限公司 | Obstacle determination method and device, electronic equipment and storage medium |
| CN113920487A (en) * | 2021-10-09 | 2022-01-11 | 广东汇天航空航天科技有限公司 | Obstacle detection processing method, device and system |
| CN113807795A (en) * | 2021-10-19 | 2021-12-17 | 上海擎朗智能科技有限公司 | Method for identifying congestion in robot distribution scene, robot and distribution system |
| CN115598656A (en) * | 2022-12-14 | 2023-01-13 | 成都运达科技股份有限公司（Cn） | Obstacle detection method, device and system based on suspension track |
| CN115598656B (en) * | 2022-12-14 | 2023-06-09 | 成都运达科技股份有限公司 | Obstacle detection method, device and system based on suspension track |
| CN115607052A (en) * | 2022-12-19 | 2023-01-17 | 科大讯飞股份有限公司 | Cleaning method, device and equipment of robot and cleaning robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109284348B (en) | Electronic map updating method, device, equipment and storage medium | |
| CN112613424A (en) | Rail obstacle detection method, rail obstacle detection device, electronic apparatus, and storage medium | |
| CN106951847B (en) | Obstacle detection method, apparatus, device and storage medium | |
| Liu et al. | A review of applications of visual inspection technology based on image processing in the railway industry | |
| CN109633688B (en) | Laser radar obstacle identification method and device | |
| CN109188438B (en) | Yaw angle determination method, device, equipment and medium | |
| CN108629231B (en) | Obstacle detection method, apparatus, device and storage medium | |
| JP2021523443A (en) | Association of lidar data and image data | |
| CN113468941B (en) | Obstacle detection method, device, equipment and computer storage medium | |
| CN108509820B (en) | Obstacle segmentation method and device, computer equipment and readable medium | |
| CN110794406B (en) | Multi-source sensor data fusion system and method | |
| CN110263713B (en) | Lane line detection method, lane line detection device, electronic device, and storage medium | |
| CN112037521B (en) | Vehicle type identification method and hazardous chemical substance vehicle identification method | |
| Ke et al. | A cost-effective framework for automated vehicle-pedestrian near-miss detection through onboard monocular vision | |
| CN110909713B (en) | Method, system and medium for extracting point cloud data track | |
| CN112519797A (en) | Vehicle safety distance early warning method, early warning system, automobile and storage medium | |
| US20220171975A1 (en) | Method for Determining a Semantic Free Space | |
| US20210350142A1 (en) | In-train positioning and indoor positioning | |
| CN113740874A (en) | Road edge detection method, electronic equipment and vehicle | |
| JP2021043146A (en) | Obstacle detection system and obstacle detection method | |
| CN113838125A (en) | Target position determining method and device, electronic equipment and storage medium | |
| CN111160132A (en) | Method and device for determining lane where obstacle is located, electronic equipment and storage medium | |
| CN114119955A (en) | Method and device for detecting potential dangerous target | |
| CN109188419B (en) | Method and device for detecting speed of obstacle, computer equipment and storage medium | |
| CN115147809A (en) | Obstacle detection method, device, equipment and storage medium |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210406 |