CN115930954A - Mining area high-precision map construction and updating method - Google Patents
Mining area high-precision map construction and updating method Download PDFInfo
- Publication number
- CN115930954A CN115930954A CN202310219106.XA CN202310219106A CN115930954A CN 115930954 A CN115930954 A CN 115930954A CN 202310219106 A CN202310219106 A CN 202310219106A CN 115930954 A CN115930954 A CN 115930954A
- Authority
- CN
- China
- Prior art keywords
- mining area
- area
- precision map
- mining
- updating
- 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.)
- Granted
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
The invention provides a mining area high-precision map construction and updating method, which comprises the following steps: collecting laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area; determining travelable region boundaries of a structured road and an unstructured operation region in a target mining area based on laser point cloud data of the target mining area and real-time pose information of vehicles travelling in the mining area to generate a mining area high-precision map of the target mining area; aiming at the mining area high-precision map, when an increment updating condition is triggered, updating the travelable area boundary of the unstructured operation area in the target mining area; and (3) carrying out quality inspection on the mining area high-precision map by taking the overall topology as a principle, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem. The method solves the technical bottleneck of high-precision map high-frequency and high-efficiency updating in the mining area operation scene.
Description
Technical Field
The invention relates to the technical field of mining area map construction and updating, in particular to a mining area high-precision map construction and updating method.
Background
With the continuous development of artificial intelligence, intelligent mines and unmanned mines become the future development trend of mining area operation. The high-precision map provides a necessary data base for realizing automatic navigation planning of the mine car. However, there is no mature solution if high-precision maps of the mine are made and maintained. The mining area high-precision map is different from a high-precision map in an urban environment. The road environment change period under the urban scene is relatively short, and the manual drawing mode is long in time consumption and high in cost, but is also a scheme capable of solving the drawing problem; in contrast, the terrain of a mining area changes very quickly, especially in an unstructured area, the terrain changes due to each loading and unloading task, and the manual drawing and maintenance method is difficult to meet the real-time requirement of map use.
Mining operation is carried out in a mining area uninterruptedly, the terrain change of a scene of the mining area is relatively quick, and the high-precision map is required to ensure the freshness and the construction and maintenance efficiency. Correspondingly, the acquisition and updating frequency of the topographic data is high, and the production efficiency of the cartographic plotting part is high. The high-frequency data acquisition is mainly realized by using a sensor equipped on an unmanned mine car for operation; however, how to efficiently process a large number of batches of collected data to maintain high-precision map data is a great challenge. The unmanned-oriented high-precision map is a vector map which accurately describes driving road surface environment and driving auxiliary information, and the main map formats which can be referred to are Opendrive, lanelet2 and the like. The existing drawing schemes are two, namely manual drawing and automatic construction. The manual drawing needs a certain knowledge reserve of a marking person for the logic topology and the field and field environment of the marked map, the vector map is drawn according to the unmanned aerial vehicle or the satellite image by means of personal experience, and the marked map is modified according to new data after the map data are updated each time; however, the manual drawing efficiency is low, errors are easy to occur, and particularly, the map maintenance result is easy to rework and reprocess due to the limited viewing angle during three-dimensional labeling, so that the drawing efficiency is reduced, and the real-time drawing requirement of a mining area is difficult to meet; the manual drawing method also needs basic business training on the labeling personnel to enable the labeling personnel to have the capability of specifically labeling the map, which undoubtedly brings higher labor cost.
The automatic construction mode generally adopts data collected by a mine card, and a vector map with certain map logic is automatically generated after a road boundary and a topological structure are extracted. For example, in the prior art (CN 109143259A), a high-precision map in an OpenDrive format is automatically generated according to GPS/IMU positioning track information acquired by a vehicle. However, the calculation result is not verified, and the unmanned system may be abnormal due to the failure of the algorithm.
Therefore, the invention provides a mining area high-precision map building and updating method.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a mining area high-precision map building and updating method, which comprises the following steps:
s1, collecting laser point cloud data of a target mining area and real-time pose information of a running vehicle in the mining area;
s2, determining travelable region boundaries of a structured road and an unstructured operation region in the target mining area based on the laser point cloud data of the target mining area and the real-time pose information of the traveling vehicles in the mining area to generate a mining area high-precision map of the target mining area;
s3, aiming at the mining area high-precision map, when an increment updating condition is triggered, updating the boundary of the travelable area of the unstructured operation area in the target mining area;
and S4, carrying out quality inspection on the mining area high-precision map by taking the overall topology as a principle, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem.
According to one embodiment of the invention, step S1 comprises:
installing a laser radar for collecting laser point cloud data and a combined navigation GPS/IMU sensor for collecting real-time pose information of a vehicle on a data collection mining vehicle;
and carrying out external parameter calibration and data hard synchronization on the laser radar and the combined navigation GPS/IMU sensor.
According to one embodiment of the invention, step S1 comprises:
the method comprises the steps that a trunk road in a target mining area is used as a structured road, a loading area in the target mining area is used as an unstructured operation area, when data are collected, a data collection mining vehicle is started on the trunk road, runs into the loading area from the trunk road, runs for a plurality of circles around the loading area, and the data collection process is finished until a running track covers all the loading area.
According to one embodiment of the invention, in step S2, the travelable zone boundary of the structured road in the target mine area is obtained by:
based on the real-time pose information of the running vehicles in the mining area, using the three-dimensional pose points of the running tracks of the vehicles as the central lines of the virtual lanes;
and determining the width of a virtual lane by taking the width of the vehicle type of the running vehicle in the mining area as a reference, and generating a boundary line of the virtual lane by combining the direction of the center line of the virtual lane to be used as the drivable area boundary of the structured road in the target mining area.
According to one embodiment of the invention, in step S2, the travelable zone boundary of the unstructured operation area in the target mine area is obtained by the following steps:
performing semantic segmentation processing on each frame of point cloud data in the laser point cloud data of the target mining area, and deducing by using a neural network model to obtain a semantic segmentation result comprising travelable road surface points and non-travelable area points;
converting laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area into the same coordinate system, and obtaining spliced point cloud data through point cloud splicing;
for an unstructured operation area, rejecting non-driving area points in the spliced point cloud data according to the semantic segmentation result;
and projecting the eliminated spliced point cloud data to a grid map with fixed resolution and three-dimensional height information, and obtaining a travelable domain boundary of the unstructured operation area in the target mining area by using a boundary detection algorithm.
According to one embodiment of the invention, in step S3, for the structured road in the target mine area, when no new cut or abandoned work area occurs, the road boundary of work area connectivity does not change, and the incremental update condition is not triggered.
According to an embodiment of the invention, in step S3, for the unstructured operation area in the target mining area, when a mining operation or a hauling operation occurs, the incremental updating condition is triggered, and the travelable area boundary updating is performed on the unstructured operation area in the target mining area through the following steps:
collecting laser point cloud data of a terrain variation area caused by mining operation or hauling operation and real-time pose information of a driving vehicle in the terrain variation area;
acquiring a feasible region closed boundary line of an unstructured operation region before mining operation or hauling operation, and recording the feasible region closed boundary line as a first feasible region closed boundary line;
acquiring a feasible region closed boundary line of the terrain variation region after mining operation or hauling operation based on laser point cloud data of the terrain variation region and real-time pose information of a running vehicle in the terrain variation region, and recording the feasible region closed boundary line as a second feasible region closed boundary line;
unifying the height values of all three-dimensional points on the first and second movable domain closed boundary lines, and respectively converting the unified height values into a first planar polygon and a second planar polygon;
if the first planar polygon does not intersect with the second planar polygon, abandoning the updating;
if the first planar polygon is intersected with the second planar polygon, merging the first planar polygon and the second planar polygon to obtain a merged planar polygon;
and attaching height values to all two-dimensional points in the combined plane polygon according to the first feasible region closed boundary line and the second feasible region closed boundary line to obtain the updated feasible driving region boundary of the unstructured operation region.
According to one embodiment of the invention, step S4 comprises:
extracting starting points and end points of the central lines of all virtual lanes in the mining area high-precision map as a starting point and end point set;
performing two-dimensionalization on all three-dimensional points in the starting and ending point set to obtain a two-dimensionalized set;
for any point in the two-dimensional set, if the point is not in a polygon formed by two-dimensional feasible region boundaries of any unstructured operation region, the mining area high-precision map has a topology disconnection problem, otherwise, the topology disconnection problem does not exist;
and if the high-precision map of the mining area has the topology disconnection problem, re-acquiring the laser point cloud data of the target mining area and the real-time pose information of the running vehicles in the mining area, and updating the high-precision map of the mining area.
According to another aspect of the invention, there is also provided a storage medium containing a series of instructions for carrying out the steps of the method as described in any one of the above.
According to another aspect of the present invention, there is also provided a mining area high-precision map building and updating apparatus, which performs the method as described in any one of the above, the apparatus comprising:
the data acquisition module is used for acquiring laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area;
the travelable region boundary extraction module is used for determining travelable region boundaries of structured roads and unstructured operation regions in the target mining area based on the laser point cloud data of the target mining area and the real-time pose information of vehicles travelling in the mining area so as to generate a mining area high-precision map of the target mining area;
the increment updating module is used for updating the boundary of the travelable domain of the unstructured operation area in the target mining area aiming at the mining area high-precision map when an increment updating condition is triggered;
and the quality inspection module is used for performing quality inspection on the mining area high-precision map on the basis of the overall topology, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem.
Compared with the prior art, the mining area high-precision map building and updating method provided by the invention has the following advantages:
1) The complete mining area high-precision map construction and updating scheme comprises the following steps: the invention provides a set of complete high-precision map construction and updating scheme from data acquisition, map construction, map updating to map review, and solves the technical bottleneck of high-precision map high-frequency and high-efficiency updating in mining operation scenes;
2) Detection of driving area boundary with strong pertinence: the invention combs two driving area boundary detection scenes of a structured road and an unstructured operation area of a mining area, provides boundary detection algorithms with strong pertinence and high robustness for the two scenes respectively, and solves the bottleneck that the mining area has no actual reference lane boundary and operation area boundary;
3) More efficient map construction and updating are realized: compared with manual drawing, the method provided by the invention has the advantages that the automation degree is high, manual participation is not needed, and the efficiency of building and updating the high-precision map of the mining area is improved on the premise of reducing the labor cost. Compared with the existing automatic labeling method, the method provided by the invention has strong robustness, and the map quality inspection examines the result of automatically constructing the map from the global perspective.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
Fig. 1 shows a flow chart of a mining area high-precision map building and updating method according to an embodiment of the invention.
FIG. 2 shows a data collection flow diagram according to one embodiment of the invention.
Fig. 3 shows a flowchart for obtaining drivable domain boundaries of a structured road in a target mine area according to an embodiment of the invention.
Fig. 4 shows a flow chart for obtaining a drivable domain boundary of an unstructured working area within a target mine area, according to an embodiment of the invention.
Fig. 5 shows a flow diagram of a condition for determining whether a structured road within a target mine area triggers an incremental update according to one embodiment of the invention.
Fig. 6 shows a flowchart of travelable zone boundary update for an unstructured work area within a target mine area, according to one embodiment of the invention.
Fig. 7 shows a quality check flow diagram according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in further detail below with reference to the accompanying drawings.
The unmanned-oriented high-precision map is a vector map which accurately describes driving road surface environment and driving auxiliary information, and the referenced mainstream map formats include Opendrive, lanelet2 and the like. The existing drawing schemes are two, namely manual drawing and automatic construction. The manual drawing needs a certain knowledge reserve of a marking person for the logic topology and the on-site and on-site environment of the marked map, the vector map is drawn according to the unmanned aerial vehicle or the satellite image by means of personal experience, and the marked map is modified according to new data after the map data is updated every time; however, the manual drawing efficiency is low, errors are easy to occur, and particularly, the map maintenance result is easy to rework and reprocess due to the limited viewing angle during three-dimensional labeling, so that the drawing efficiency is reduced, and the real-time drawing requirement of a mining area is difficult to meet; the manual drawing method also needs basic business training on the labeling personnel to enable the labeling personnel to have the capability of specifically labeling the map, which undoubtedly brings higher labor cost.
The automatic construction mode generally adopts data collected by a mine card, and a vector map with certain map logic is automatically generated after a road boundary and a topological structure are extracted. For example, in the prior art (CN 109143259A), a high-precision map in an OpenDrive format is automatically generated according to GPS/IMU positioning track information acquired by a vehicle. However, the calculation result is not verified, and the unmanned system may be abnormal because of the failure of the algorithm.
Moreover, the existing automatic construction method lacks of quality check of the map. The automatic construction method is often designed for certain single scenes, so that the condition that the construction algorithm is invalid or abnormal values occur is difficult to avoid in actual production.
Aiming at the defects in the prior art, the invention provides a high-precision map construction and updating method with high speed, high efficiency and strong reliability aiming at mining area scenes, and the invention is provided with an inspection mechanism, can inspect whether the automatically constructed map conforms to the map structure logic or not, and improves the reliability of map data.
Fig. 1 shows a flow chart of a mining area high-precision map building and updating method according to an embodiment of the invention.
As shown in fig. 1, in step S1, laser point cloud data of a target mine area and real-time pose information of a vehicle running in the mine area are collected.
In one embodiment, as shown in fig. 2, step S1 includes step S201 of installing a lidar for acquiring laser point cloud data and a combined navigation GPS/IMU sensor for acquiring real-time pose information of the vehicle on the data acquisition mining vehicle. In particular, mining vehicles include, but are not limited to, mining trucks on which lidar and combined navigation GPS/IMU sensors are mounted as data acquisition mining vehicles. The combined navigation GPS/IMU sensor includes a GPS (Global Positioning System) sensor and an IMU (Inertial Measurement Unit) sensor.
In one embodiment, as shown in FIG. 2, step S1 includes step S202 of performing extrinsic parameter calibration and hard data synchronization on the lidar and the integrated navigation GPS/IMU sensors. Specifically, a laser radar and a combined navigation GPS/IMU sensor are mounted on a mining truck, external parameter calibration and data hard synchronization are carried out on the two devices, the mining truck is acquired through data, the mining truck runs along a normal operation path, the laser radar acquires laser point cloud data in the running process, and the combined navigation GPS/IMU sensor acquires data and acquires real-time pose information of the mining truck.
In one embodiment, as shown in fig. 2, step S1 includes step S203, taking the main road in the target mining area as a structured road, taking the loading area in the target mining area as an unstructured operation area, when data acquisition is performed, starting the data acquisition mining vehicle on the main road, driving the mining vehicle from the main road into the loading area, and driving the mining vehicle around the loading area for several turns until the driving track covers all the loading area, and then ending the data acquisition process. Specifically, the trunk road is connected with each loading area, the data acquisition mining vehicle is started on the trunk road, the trunk road runs into the loading area, the data acquisition mining vehicle runs for a plurality of circles around the loading area until the track covers all loading areas, the data acquisition is finished, and the industrial personal computer of the data acquisition mining vehicle continuously acquires laser point cloud data generated by a laser radar and positioning data generated by a combined navigation GPS/IMU sensor in the process.
As shown in fig. 1, in step S2, travelable zone boundaries of structured roads and unstructured operation areas in the target mine area are determined based on the laser point cloud data of the target mine area and real-time pose information of vehicles traveling in the mine area, so as to generate a mine area high-precision map of the target mine area. In particular, the travelable zones of a mine may be broadly divided into two categories: structured roads and unstructured work areas.
In one embodiment, the travelable zone boundary of the structured road in the target mine is obtained through the steps shown in fig. 3, and in step S301, the three-dimensional pose point of the vehicle travel track is used as the virtual lane center line based on the real-time pose information of the traveling vehicle in the mine. Specifically, for the structured road boundary, since the actual mine road environment does not have the lane line identification like an urban road, the three-dimensional pose point of the vehicle driving track is required to be used as the virtual lane center line.
In one embodiment, as shown in fig. 3, in step S302, a virtual lane width is determined based on the vehicle type width of the running vehicle in the mine area, and a virtual lane boundary line is generated in combination with the virtual lane center line direction as the travelable zone boundary of the structured road in the target mine area. Specifically, the vehicle type having the widest vehicle width among the types of vehicles which travel more frequently on the mine road is selected, and the length 1.5 times as long as the vehicle type width is used as the virtual lane width, so that the virtual lane boundary line is generated according to the virtual lane center line direction.
In one embodiment, for an unstructured work area, such an area is a free-driving area within the boundaries of a drivable domain, within which there is no logically defined "road". The boundary of the region of the type can be obtained through several calculation steps of point cloud semantic segmentation, point cloud splicing and boundary detection.
Further, the travelable region boundary of the unstructured operation region in the target mine area is obtained through the steps shown in fig. 4, in step S401, semantic segmentation processing is performed on each frame of point cloud data in the laser point cloud data of the target mine area, and a semantic segmentation result including travelable road surface points and non-travelable region points is obtained through inference by using a neural network model. Specifically, semantic segmentation processing is carried out on each frame of point cloud, and two classification points, namely a drivable road surface point and an undrivable area point, are obtained through inference by using a randet + + neural network model. The randent + + neural network model converts the laser point cloud into distance Images (Range Images) through spherical projection, and then performs semantic segmentation on the distance Images by using two-dimensional convolutional neural network extraction features. In order to obtain an accurate segmentation effect, a post-processing algorithm can be adopted to process a discretization error caused by projection transformation or an output result blurred by a convolutional neural network.
In one embodiment, as shown in fig. 4, in step S402, the laser point cloud data of the target mine area and the real-time pose information of the driving vehicle in the mine area are converted into the same coordinate system, and the joined point cloud data is obtained through point cloud joining. Specifically, for a hard synchronous combined navigation GPS/IMU sensor and a laser radar sensor, laser point cloud data can be converted from a laser radar coordinate SYSTEM to a unified UTM (UNIVERSAL TRANSVERSE graphite machine GRID SYSTEM) coordinate SYSTEM by external reference and GPS/IMU positioning data of the two sensors, and point cloud splicing is realized based on the unified coordinate SYSTEM.
In one embodiment, as shown in fig. 4, in step S403, for an unstructured operation area, an undriven area point in the stitched point cloud data is removed according to a semantic segmentation result. Specifically, the spliced local point cloud covers the whole unstructured operation area, and the points of the non-drivable area of the local point cloud are removed according to the result of semantic segmentation.
In one embodiment, as shown in fig. 4, in step S404, the eliminated joined point cloud data is projected to a grid map with fixed resolution and three-dimensional height information, and a travelable region boundary of an unstructured operation region in the target mine area is obtained by using a boundary detection algorithm. Specifically, the eliminated spliced point cloud data is projected to a grid map with fixed resolution and three-dimensional height information, and a boundary detection algorithm based on a canny operator (canny edge detection operator) is utilized to obtain a point cloud boundary of the travelable area.
As shown in fig. 1, in step S3, when an incremental update condition is triggered for the mine area high-precision map, a travelable area boundary update is performed for the unstructured work area within the target mine area.
In the actual mining operation, in step S3, if no new operating area or abandoned operating area is found, the road boundary communicated between the unstructured operating areas will not change, so that the structured road boundary does not need to be updated incrementally.
In one embodiment, as shown in fig. 5, in step 501 and step 502, for the structured road in the target mining area, when no new or abandoned working area is found, the road boundary of the working area connection is not changed, and the increment updating condition is not triggered.
For the structured roads in the target mining area, the road structure cannot be changed frequently, the road boundary cannot be changed, and the increment updating condition is not triggered. Further, if a newly opened operation area or a waste operation area occurs and the road structure changes, the map reconstruction should be triggered, refer to steps S301 to S302, and the incremental update condition is not triggered. When a newly opened operation area appears, a driver drives the data acquisition mining vehicle to acquire data so as to increase a road communicated with the newly opened operation area in a high-precision map of a mining area.
For unstructured work areas, the travelable zone boundaries are altered every mining and hauling operation. Therefore, for the unstructured operation area in the target mining area, when mining operation or hauling operation occurs, an incremental updating condition is triggered, the travelable area boundary of the unstructured operation area in the target mining area is updated through the steps shown in fig. 6, and in step S601, laser point cloud data of a terrain variation area caused by the mining operation or hauling operation and real-time pose information of a vehicle traveling in the terrain variation area are collected. Specifically, laser point cloud data of a terrain variation area caused by a mining operation or a hauling operation and real-time pose information of a running vehicle in the terrain variation area are collected through the steps shown in fig. 2 and/or fig. 3. Furthermore, after the loading area is subjected to mining and hauling operation for a period of time, the data acquisition mining vehicle is used for acquiring topographic data of which the loading area only has topographic changes until the track covers all the changed parts, and then the data acquisition is finished, and in the process, the industrial control computer continuously acquires laser point cloud data generated by the laser radar and positioning data generated by the combined navigation GPS/IMU sensor.
In one embodiment, as shown in fig. 6, in step S602, a feasible region closed boundary line of the unstructured operation region before the mining operation or the hauling operation is acquired and recorded as a first feasible region closed boundary line. Specifically, a feasible region closed boundary line of the unstructured operation region before the mining operation or the hauling operation is acquired by using a feasible region boundary extraction algorithm and recorded as a first feasible region closed boundary line E1.
In one embodiment, as shown in fig. 6, in step S603, a feasible region closed boundary line of the terrain variation area after the mining operation or the hauling operation is acquired and recorded as a second feasible region closed boundary line based on the laser point cloud data of the terrain variation area and the real-time pose information of the traveling vehicle in the terrain variation area. Specifically, based on the laser point cloud data of the terrain variation area and the real-time pose information of the driving vehicle in the terrain variation area, the travelable domain boundary of the terrain variation area after mining operation or hauling operation is obtained through the steps shown in fig. 5, and the travelable domain closed boundary line of the terrain variation area after mining operation or hauling operation is obtained by using a travelable domain boundary extraction algorithm and recorded as a second travelable domain closed boundary line Δ E.
In one embodiment, as shown in fig. 6, in step S604, the height values of all three-dimensional points on the first and second horizontal closed boundary lines are uniformly converted into a first and second planar polygon, respectively. Specifically, the height values of all three-dimensional points on E1 and Δ E are unified and converted into a first planar polygon S1 and a second planar polygon Δ S, respectively.
In one embodiment, as shown in fig. 6, in step S605, if the first planar polygon does not intersect with the second planar polygon, the update is discarded. Specifically, a geometric calculation method is used to determine whether S1 and Δ S intersect, and if S1 and Δ S do not intersect, the update is discarded.
In one embodiment, as shown in fig. 6, in step S606, if the first planar polygon intersects with the second planar polygon, the first planar polygon and the second planar polygon are merged to obtain a merged planar polygon. Specifically, whether S1 and Δ S intersect is determined by using a computational geometry method, and if S1 and Δ S intersect, the planar polygons S2 = S1 uberΔ S are merged.
In one embodiment, as shown in fig. 6, in step S607, height values are attached to all two-dimensional points in the merged planar polygon according to the first feasible region closed boundary line and the second feasible region closed boundary line, so as to obtain the feasible region boundary of the updated unstructured operation region. Specifically, the height values are attached to all the two-dimensional points in S2 according to E1 and Δ E, and the travelable region boundary E2 of the updated unstructured operation region is obtained.
As shown in fig. 1, in step S4, the quality of the mine area high-precision map is checked based on the overall topology, and when the mine area high-precision map has a topology disconnection problem, the mine area high-precision map is updated.
Because the steps S1-S3 are divided into two types of data of the structured road and the unstructured operation area for separate discussion, the overall topology of the high-precision map is not considered. The map data are qualified to be issued to each unmanned terminal for application only when the map updating of each time passes topology check, otherwise 'furnace return remanufacturing' is needed, and updating data need to be collected again to correct and legally modify the map. The step S4 provided by the invention ensures that the elements in the map are legal logically from the structure, and improves the reliability of the map data.
In one embodiment, the quality of the mine area high-precision map is checked through the steps shown in fig. 7, and in step S701, the starting points and the end points of all the virtual lane center lines in the mine area high-precision map are extracted as the starting and end point set. Specifically, the set P is a set of three-dimensional points of the start point and the end point of the virtual lane center lines of all the lanes.
In one embodiment, as shown in fig. 7, in step S702, all three-dimensional points in the starting and ending point sets are subjected to two-dimensioning to obtain a two-dimensioned set. Specifically, all points in the set P are subjected to two-dimensionalization to obtain a two-dimensionalized set P'. Further, unifying the height values of all three-dimensional points in the set P to obtain a two-dimensional set P'.
In one embodiment, as shown in fig. 7, in step S703, if any point in the two-dimensional set is not located in a polygon formed by two-dimensional feasible region boundaries of any unstructured working region, the high-precision map of the mining area has a topology disconnected problem, otherwise, the topology disconnected problem does not exist. Specifically, for any point P ' in P ', if the point P ' is not in the polygon S formed by the two-dimensional feasible region boundaries of any unstructured operation region, the map topology has a non-connection problem, otherwise, the map topology does not have a topology non-connection problem.
In one embodiment, as shown in fig. 7, in step S704, if the mine area high-precision map has a topology disconnection problem, the laser point cloud data of the target mine area and the real-time pose information of the driving vehicles in the mine area are collected again, so as to update the mine area high-precision map. Specifically, based on the laser point cloud data of the target mine area obtained by re-acquisition and the real-time pose information of the driving vehicle in the mine area, the driving area boundary of the structured road and the non-structured operation area in the target mine area are respectively obtained through the steps shown in fig. 4 and 5, and the high-precision map of the mine area is updated. And if the mining area high-precision map does not have the topology disconnection problem, directly publishing the new version of map data after incremental updating in the step S3, or directly publishing the mining area high-precision map of the target mining area generated in the step S2.
The mining area high-precision map building and updating method provided by the invention can also be matched with a computer readable storage medium, and a computer program is stored on the storage medium and is executed to operate the mining area high-precision map building and updating method. The computer program is capable of executing computer instructions comprising computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc.
The computer-readable storage medium may include: any entity or device capable of carrying computer program code, recording medium, U.S. disk, removable hard disk, magnetic diskette, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signal, telecommunications signal, software distribution medium, etc.
According to another aspect of the invention, there is also provided a mining area high-precision map building and updating device, which executes a mining area high-precision map building and updating method, and the device comprises: the device comprises a data acquisition module, a travelable region boundary extraction module, an increment updating module and a quality inspection module.
The data acquisition module is used for acquiring laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area; the travelable region boundary extraction module is used for determining travelable region boundaries of structured roads and unstructured operation regions in the target mining area based on laser point cloud data of the target mining area and real-time pose information of vehicles travelling in the mining area so as to generate a mining area high-precision map of the target mining area; the increment updating module is used for updating the boundary of the travelable domain of the unstructured operation area in the target mining area aiming at the mining area high-precision map when an increment updating condition is triggered; the quality inspection module is used for inspecting the quality of the mining area high-precision map on the basis of the overall topology, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem.
In summary, compared with the prior art, the mining area high-precision map construction and updating method provided by the invention has the following advantages:
1) The complete mining area high-precision map construction and updating scheme comprises the following steps: the invention provides a set of complete high-precision map construction and updating scheme from data acquisition, map construction, map updating to map review, and solves the technical bottleneck of high-frequency and high-efficiency updating of high-precision maps in mining area operation scenes;
2) And (3) detecting the boundary of the drivable area with strong pertinence: the invention combs two driving area boundary detection scenes of a structured road and an unstructured operation area of a mining area, provides boundary detection algorithms with strong pertinence and high robustness for the two scenes respectively, and solves the bottleneck that the mining area has no actual reference lane boundary and operation area boundary;
3) More efficient map construction and updating are realized: compared with manual drawing, the method provided by the invention has the advantages that the automation degree is high, manual participation is not needed, and the efficiency of building and updating the high-precision map of the mining area is improved on the premise of reducing the labor cost. Compared with the existing automatic labeling method, the method provided by the invention has strong robustness, and the map quality inspection examines the result of automatically constructing the map from the global perspective.
Claims (10)
1. A mining area high-precision map building and updating method, characterized in that the method comprises:
s1, collecting laser point cloud data of a target mining area and real-time pose information of a running vehicle in the mining area;
s2, determining travelable region boundaries of a structured road and an unstructured operation region in the target mining area based on the laser point cloud data of the target mining area and the real-time pose information of the traveling vehicles in the mining area to generate a mining area high-precision map of the target mining area;
s3, aiming at the mining area high-precision map, when an increment updating condition is triggered, updating the boundary of the travelable area of the unstructured operation area in the target mining area;
and S4, carrying out quality inspection on the mining area high-precision map by taking the overall topology as a principle, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem.
2. The mining area high-precision map building and updating method according to claim 1, wherein the step S1 comprises the following steps:
installing a laser radar for collecting laser point cloud data and a combined navigation GPS/IMU sensor for collecting real-time pose information of a vehicle on a data collection mining vehicle;
and carrying out external parameter calibration and data hard synchronization on the laser radar and the combined navigation GPS/IMU sensor.
3. The mining area high-precision map building and updating method according to claim 2, wherein the step S1 comprises the following steps:
the method comprises the steps that a trunk road in a target mining area is used as a structured road, a loading area in the target mining area is used as an unstructured operation area, when data are collected, a data collection mining vehicle is started on the trunk road, runs into the loading area from the trunk road, runs for a plurality of circles around the loading area, and the data collection process is finished until a running track covers all the loading area.
4. The mining area high-precision map building and updating method according to claim 1, wherein in the step S2, the drivable-range boundary of the structured road in the target mining area is obtained through the following steps:
based on the real-time pose information of the running vehicles in the mining area, using the three-dimensional pose points of the running tracks of the vehicles as the central lines of the virtual lanes;
and determining the width of a virtual lane by taking the width of the vehicle type of the running vehicle in the mining area as a reference, and generating a boundary line of the virtual lane by combining the direction of the center line of the virtual lane to be used as the drivable area boundary of the structured road in the target mining area.
5. The mining area high-precision map building and updating method according to claim 1, wherein in step S2, the travelable area boundary of the unstructured operation area in the target mining area is obtained through the following steps:
performing semantic segmentation processing on each frame of point cloud data in the laser point cloud data of the target mining area, and deducing by using a neural network model to obtain a semantic segmentation result containing travelable road points and non-travelable area points;
converting laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area into the same coordinate system, and obtaining spliced point cloud data through point cloud splicing;
for an unstructured operation area, rejecting non-driving area points in the spliced point cloud data according to the semantic segmentation result;
and projecting the eliminated spliced point cloud data to a grid map with fixed resolution and three-dimensional height information, and obtaining a travelable domain boundary of the unstructured operation area in the target mining area by using a boundary detection algorithm.
6. A mining area high-precision map building and updating method according to claim 1, wherein in step S3, for the structured road in the target mining area, when no new excavation operation area or waste operation area appears, the road boundary of the working area connection does not change, and the increment updating condition is not triggered.
7. The mining area high-precision map building and updating method according to claim 1, wherein in step S3, for the unstructured operation area in the target mining area, when mining operation or hauling operation occurs, the incremental updating condition is triggered, and the travelable area boundary updating is carried out on the unstructured operation area in the target mining area through the following steps:
collecting laser point cloud data of a terrain variation area caused by mining operation or hauling operation and real-time pose information of a driving vehicle in the terrain variation area;
acquiring a feasible region closed boundary line of an unstructured operation region before mining operation or hauling operation, and recording the feasible region closed boundary line as a first feasible region closed boundary line;
acquiring a feasible region closed boundary line of the terrain variation region after mining operation or hauling operation based on laser point cloud data of the terrain variation region and real-time pose information of a running vehicle in the terrain variation region, and recording the feasible region closed boundary line as a second feasible region closed boundary line;
unifying the height values of all three-dimensional points on the first and second movable domain closed boundary lines, and respectively converting the unified height values into a first planar polygon and a second planar polygon;
if the first planar polygon does not intersect with the second planar polygon, discarding the update;
if the first planar polygon is intersected with the second planar polygon, merging the first planar polygon with the second planar polygon to obtain a merged planar polygon;
and attaching height values to all two-dimensional points in the combined plane polygon according to the first feasible region closed boundary line and the second feasible region closed boundary line to obtain the updated feasible driving region boundary of the unstructured operation region.
8. A mining area high-precision map building and updating method according to any one of claims 1-7, wherein the step S4 comprises:
extracting starting points and end points of the central lines of all virtual lanes in the mining area high-precision map as a starting point and end point set;
performing two-dimensionization on all three-dimensional points in the starting and ending point set to obtain a two-dimensionized set;
for any point in the two-dimensional set, if the point is not in a polygon formed by two-dimensional feasible region boundaries of any unstructured operation region, the mining area high-precision map has a topology disconnection problem, otherwise, the topology disconnection problem does not exist;
and if the high-precision map of the mining area has the topology disconnection problem, re-acquiring the laser point cloud data of the target mining area and the real-time pose information of the running vehicles in the mining area, and updating the high-precision map of the mining area.
9. A storage medium characterized in that it contains a series of instructions for carrying out the steps of the method according to any one of claims 1 to 8.
10. A mine high-precision map building and updating device, characterized by executing the method according to any one of claims 1-8, wherein the device comprises:
the data acquisition module is used for acquiring laser point cloud data of a target mining area and real-time pose information of vehicles running in the mining area;
the travelable region boundary extraction module is used for determining travelable region boundaries of structured roads and unstructured operation regions in the target mining area based on the laser point cloud data of the target mining area and the real-time pose information of vehicles travelling in the mining area so as to generate a mining area high-precision map of the target mining area;
the increment updating module is used for updating the travelable domain boundary of the unstructured operation area in the target mining area aiming at the mining area high-precision map when an increment updating condition is triggered;
and the quality inspection module is used for carrying out quality inspection on the mining area high-precision map on the basis of the overall topology, and updating the mining area high-precision map when the mining area high-precision map has the topology disconnection problem.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310219106.XA CN115930954B (en) | 2023-03-09 | 2023-03-09 | Mining area high-precision map construction and updating method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310219106.XA CN115930954B (en) | 2023-03-09 | 2023-03-09 | Mining area high-precision map construction and updating method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115930954A true CN115930954A (en) | 2023-04-07 |
| CN115930954B CN115930954B (en) | 2023-06-30 |
Family
ID=86650982
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310219106.XA Active CN115930954B (en) | 2023-03-09 | 2023-03-09 | Mining area high-precision map construction and updating method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115930954B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116753942A (en) * | 2023-08-16 | 2023-09-15 | 北京路凯智行科技有限公司 | Hierarchical updated mining area high-precision map manufacturing method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111125135A (en) * | 2019-12-19 | 2020-05-08 | 北京易控智驾科技有限公司 | Map updating method, map fusion updating method, terminal and server |
| US20200225044A1 (en) * | 2017-10-05 | 2020-07-16 | Toyota Jidosha Kabushiki Kaisha | Map information provision system |
| CN111551958A (en) * | 2020-04-28 | 2020-08-18 | 北京踏歌智行科技有限公司 | Mining area unmanned high-precision map manufacturing method |
| CN111780768A (en) * | 2020-07-20 | 2020-10-16 | 北京易控智驾科技有限公司 | Loading position area map updating method applied to automatic driving of surface mine |
| WO2021171560A1 (en) * | 2020-02-28 | 2021-09-02 | 三菱電機株式会社 | High-accuracy map database generation device and high-accuracy map database generation method |
| US20210278241A1 (en) * | 2020-03-09 | 2021-09-09 | Toyota Jidosha Kabushiki Kaisha | Map data updating device and storage medium storing computer program for map data updating |
| WO2022000260A1 (en) * | 2020-06-30 | 2022-01-06 | 深圳市大疆创新科技有限公司 | Map updating method and apparatus, and movable platform and storage medium |
| CN115423958A (en) * | 2022-09-01 | 2022-12-02 | 北京踏歌智行科技有限公司 | Mining area travelable area boundary updating method based on visual three-dimensional reconstruction |
-
2023
- 2023-03-09 CN CN202310219106.XA patent/CN115930954B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200225044A1 (en) * | 2017-10-05 | 2020-07-16 | Toyota Jidosha Kabushiki Kaisha | Map information provision system |
| CN111125135A (en) * | 2019-12-19 | 2020-05-08 | 北京易控智驾科技有限公司 | Map updating method, map fusion updating method, terminal and server |
| WO2021171560A1 (en) * | 2020-02-28 | 2021-09-02 | 三菱電機株式会社 | High-accuracy map database generation device and high-accuracy map database generation method |
| US20210278241A1 (en) * | 2020-03-09 | 2021-09-09 | Toyota Jidosha Kabushiki Kaisha | Map data updating device and storage medium storing computer program for map data updating |
| CN111551958A (en) * | 2020-04-28 | 2020-08-18 | 北京踏歌智行科技有限公司 | Mining area unmanned high-precision map manufacturing method |
| WO2022000260A1 (en) * | 2020-06-30 | 2022-01-06 | 深圳市大疆创新科技有限公司 | Map updating method and apparatus, and movable platform and storage medium |
| CN111780768A (en) * | 2020-07-20 | 2020-10-16 | 北京易控智驾科技有限公司 | Loading position area map updating method applied to automatic driving of surface mine |
| CN115423958A (en) * | 2022-09-01 | 2022-12-02 | 北京踏歌智行科技有限公司 | Mining area travelable area boundary updating method based on visual three-dimensional reconstruction |
Non-Patent Citations (1)
| Title |
|---|
| 李浩然等: "基于车云协同的露天矿山数字地图更新方法", 控制与信息技术 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116753942A (en) * | 2023-08-16 | 2023-09-15 | 北京路凯智行科技有限公司 | Hierarchical updated mining area high-precision map manufacturing method |
| CN116753942B (en) * | 2023-08-16 | 2023-11-24 | 北京路凯智行科技有限公司 | Hierarchical updated mining area high-precision map manufacturing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115930954B (en) | 2023-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111551958B (en) | Mining area unmanned high-precision map manufacturing method | |
| CN109470254B (en) | Map lane line generation method, device, system and storage medium | |
| CN112069856A (en) | Map generation method, driving control method, device, electronic equipment and system | |
| AU2021313775A1 (en) | Drivable surface identification techniques | |
| US20210001891A1 (en) | Training data generation for dynamic objects using high definition map data | |
| CN110780305A (en) | Track cone bucket detection and target point tracking method based on multi-line laser radar | |
| CN110850439A (en) | High-precision three-dimensional point cloud map construction method | |
| CN105667518A (en) | Lane detection method and device | |
| CN112432650B (en) | High-precision map data acquisition method, vehicle control method and device | |
| CN115930954A (en) | Mining area high-precision map construction and updating method | |
| CN111681172A (en) | Method, equipment and system for cooperatively constructing point cloud map | |
| CN110458080A (en) | The pre-judging method and system of front pit-hole in a kind of running car | |
| Bao et al. | High-definition map generation technologies for autonomous driving | |
| CN115375866A (en) | Method, device, equipment and medium for updating three-dimensional geological model of mining area | |
| CN115439621A (en) | Three-dimensional map reconstruction and target detection method for coal mine underground inspection robot | |
| CN114359494A (en) | Mining area three-dimensional map collecting and updating method and system | |
| Farrell et al. | Best practices for surveying and mapping roadways and intersections for connected vehicle applications | |
| CN112633812A (en) | Track segmentation method, device, equipment and storage medium for freight vehicle | |
| CN115511901B (en) | Mine map processing method, device and system, computer equipment and storage medium | |
| CN113392170A (en) | High-precision map data generation method and device and server | |
| CN115375861A (en) | Three-dimensional mapping method, device and storage medium for unmanned mining area | |
| CN112513876B (en) | Road surface extraction method and device for map | |
| US20240094027A1 (en) | Method and apparatus for incremental mapping of haul roads | |
| Li et al. | A Ground Segmentation Method Based on Point Cloud Map for Unstructured Roads | |
| CN114216469B (en) | Method for updating high-precision map, intelligent base station 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |