CN114791936A - Storage, efficient editing and calling method for passable area of unmanned vehicle - Google Patents
Storage, efficient editing and calling method for passable area of unmanned vehicle Download PDFInfo
- Publication number
- CN114791936A CN114791936A CN202110101983.8A CN202110101983A CN114791936A CN 114791936 A CN114791936 A CN 114791936A CN 202110101983 A CN202110101983 A CN 202110101983A CN 114791936 A CN114791936 A CN 114791936A
- Authority
- CN
- China
- Prior art keywords
- map
- sub
- maps
- vehicle
- 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
- 238000000034 method Methods 0.000 title claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 10
- 230000004927 fusion Effects 0.000 claims description 2
- 230000000877 morphologic effect Effects 0.000 claims 1
- 230000007613 environmental effect Effects 0.000 abstract description 2
- 230000004888 barrier function Effects 0.000 description 5
- 238000001514 detection method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
- G06F16/2291—User-Defined Types; Storage management thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/001—Texturing; Colouring; Generation of texture or colour
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Software Systems (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention relates to a method for storing, efficiently editing and calling a passable area of an unmanned vehicle. The method comprises the steps of merging, editing and splitting a two-dimensional grid map formed after laser radar point cloud processing. The combined global map is edited, so that the editing efficiency can be effectively improved, dynamic obstacles can be removed through editing, the accuracy and precision of boundary constraint and passable areas are enhanced, and then the combined global map is disassembled into sub-maps for path planning. The edited sub-map can better reflect the environmental information, thereby providing a basis for the accurate path planning of the vehicle safety. The method can effectively reduce the condition that the path planning is deviated or unreasonable due to the sparse boundary characteristics.
Description
Technical Field
The invention relates to the technical field of unmanned vehicles, in particular to a laser radar grid map-based acquisition and image processing technology.
Background
At present, a three-dimensional laser radar is mainly used for creating a map, and at present, the main map creating modes based on the three-dimensional laser radar comprise a feature map, a point cloud map, a topological map, a grid map, a mixed map and the like. The grid map comprises a two-dimensional grid map and a three-dimensional grid map, the grid map performs sparsification on the original point cloud by setting grid resolution, the calculation complexity is greatly reduced, numerical values in a [0, 1] interval are generally stored in the grid to represent the probability of the occupied grid, and the robustness of the grid map to system noise can be enhanced by updating the probability of the occupied grid. By dividing the grid state, the grid map can be easily used for unmanned vehicle navigation, and Konrad et al updates the grid data in one or several sub-maps at the current moment by dividing the grid map into a plurality of sub-maps, so that the memory occupation amount and the data amount of the grid map data can be controlled according to the actual computing capacity of the computing platform, and the problems of the increase of the memory occupation amount and the computing amount of a computer caused by the increase of the grid map data are avoided. The method comprises the steps of detecting obstacles by using a three-dimensional laser radar to obtain an obstacle map comprising positive obstacles, negative obstacles, cliffs, slopes and dynamic obstacles, performing multi-map fusion according to the size and the pose of a map provided by a quasi-passable area map sent by a laser radar odometer, performing Bayesian probability updating and map expansion processing on the fused map to obtain a final passable area map, and taking the extracted passable area map as a navigation map of the unmanned vehicle for planning local paths of the unmanned vehicle.
Disclosure of Invention
In view of the above analysis, we propose a method for storing, efficiently editing and invoking a passable area for unmanned vehicles. The method comprises the steps of merging, editing and splitting the extracted passable area sub-map. The combined global map is edited to effectively improve the editing efficiency, environmental noise, observation noise and pose estimation noise can be filtered out through editing, dynamic obstacles are removed, the boundary constraint and the accuracy and precision of a passable area are enhanced, and then the combined global map is disassembled into sub-maps for path planning. The edited sub-map can better optimize the path planning of the unmanned vehicle, and can effectively reduce the condition that the path planning is deviated or unreasonable due to sparse boundary features.
The method for realizing the storage, efficient editing and calling technology of the accessible area of the unmanned vehicle mainly comprises the following steps:
A. passable region extraction
The passable area is mainly divided into three steps:
first, receiving messages of a plurality of different maps and messages of an odometer, wherein the messages comprise: a quasi-passable area map, a vehicle pose, a mileometer pose, a positive and negative obstacle map, a cliff map and a slope map. And performing pose transformation on the three maps, namely the positive and negative obstacle map, the cliff map and the slope map which are transmitted in real time by taking the vehicle body as the origin of a coordinate system according to the pose of the odometer and the pose of the vehicle, and fusing the three maps into a map of a quasi passable area to obtain a passable area which is not processed yet.
After that, the obtained map is subjected to the processing based on the kernel dilation and then subjected to the bayesian probability updating, and it is noted that the probability in the area in front of the vehicle is not updated for the false detection of positive and negative obstacles. Note that this probability does not update the obstacle, but does not update the obstacle of the present frame. If the obstacle is detected in the previous frame, the probability is not updated in the limited area, and the area is still the obstacle. This has a filtering effect on the false detection of obstacles suddenly flashing near the vehicle during movement, but this can be detected when the vehicle is stationary, such as when the person walks from the vehicle to the front of the vehicle, and the purpose is mainly to eliminate the false detection of obstacles flashing when the vehicle is moving. Finally, the obtained probability map is converted into an occupation grid map, wherein 2 represents an obstacle, 1 represents passable and 0 represents unknown. And performing inflation processing based on the semi-vehicle width on the occupancy grid map to obtain a final passable map.
Describing a Bayesian probability updating algorithm principle:
based on the principle that the voxels with a larger number of laser concentration times are occupied with a higher degree of certainty and a higher occupation probability, the occupation probability of each voxel is updated by the equations 1.1 and 1.2.
M new (x)＝clamp(odds -1 (odds(M old (x))·odds(p hit )))
Where p represents the initial probability value of an occupied or empty grid, M new (x) For normalized probability values, odds (p) is the conversion of probability values to odds values, and odds ^ -1 is the conversion of odds values to probability values.
B. Sub-map stitching
After the passable area is extracted, the generated file comprises a series of sub-maps and texts capable of reading the position and pose information of the vehicle in each sub-map. In order to enhance the boundary constraint of the passable area and remove the interference of the dynamic obstacles, the map needs to be edited to achieve the expected effect. However, the editing efficiency of a single picture is low, so that all the sub-pictures are combined into a complete global map for editing through an image processing technology, and the pose points of the vehicles are displayed on the spliced global map during editing.
The image splicing specific process comprises the following steps: the total number of the sub-maps, the initial pose information of the sub-maps, the number of the current sub-maps, the resolution of the sub-maps, the height and width of the pictures, the pose of the sub-maps, the pixel coordinates of the vehicles in the sub-maps and the pose information of the pixel coordinates in the world coordinate system, which are output in the accessible area extraction and storage process. And taking the first sub-map of the accessed passable area as the original point of the splicing of all the maps, taking the pose information of the vehicles in the map as the reference, and converting and unifying the vehicle poses of all other sub-maps to the first sub-map through a rotation matrix and a translation matrix, wherein the coordinate system of the sub-maps is equivalent to the vehicle body coordinate system in the first sub-map. The method comprises the steps of obtaining relevant pose information from a text stored in a passable region, obtaining a rotation matrix according to the pose information of each sub-map and a first sub-map serving as an origin, obtaining a rotation matrix and a translation matrix on the basis of the pose of an image center for convenient calculation because the pixel indexes of vehicles in each sub-map are different, and obtaining the pose of each point in the whole sub-map according to the resolution of the map because the pixel positions and the poses of the vehicles are known.
And when the sub-maps rotate, central rotation is adopted, the image centers of the sub-maps are used as rotation centers, then the rotation angle and the translation distance are calculated, and all the sub-maps are spliced according to the obtained rotation angle and translation distance. The principle of splicing is as follows: if there is an obstacle in any one of the images, there is an obstacle in the stitched image. In the continuous splicing process, the original point information of the original first map is updated, the splicing basis is the original first sub map, in the splicing process, a lot of information is supplemented around the original map, and if the position of the original map in the new map cannot be obtained, the splicing basis is lost, so the position of the original map in the continuously expanded new map needs to be updated. Here the upper left corner of the image is used as a position marker. Therefore, the information of the updated origin needs to be calculated, and for later map disassembly, the position of each sub-map in the spliced global map needs to be recorded, and the main recording contents include: rotation angle, length and width after rotation, and position information of the rotated sub-map in the new map.
Also, as mentioned above, since the sub-map needs to be rotated continuously during the stitching process, the length and width of the rotated image are changed in order to ensure that the original image information is not lost after the rotation. Because each sub-map is different in size, some sub-maps are very wide, and some sub-maps may be very long, after the sub-maps are rotated and translated, the length and the width of the new map after splicing cannot be simply represented, and it is required to judge that the splicing position is approximately four parts, namely, upper left, lower left, upper right and lower right, according to the translation relationship between the sub-map to be spliced and the first map, and calculate the length and the width of the new map according to the length and the width information of the rotated sub-map, the splicing position and the length and the width information of the existing map.
The whole process can be realized by using a correlation function in an opencv library, for example, a copymakebox function is used for expanding the boundary of a picture, the pixel is filled in the boundary, a getrotontionmatrix 2D function rotates the picture to obtain a rotation matrix and scales the rotation matrix, affine transformation is performed by using a warpAffine function finally, the posture of the picture is changed, splicing is performed according to related information, a complete global map is spliced, the map contains information of all sub-maps and is represented as a complete map of an area where a vehicle runs, and then the whole map is edited according to needs.
C. Map editing
After the sub-maps are spliced, the spliced map needs to be edited so as to remove dynamic obstacles in the map, meanwhile, the boundary constraint of the map is enhanced as required, and the requirement for optimizing the local path planning can be met. The method for editing the pictures can be realized by manually editing the pictures, and the traces of the dynamic obstacles are manually removed by manually using related picture editing software, because the historical frame data are continuously superposed, the moving objects are continuously detected at different positions, so that a moving track is formed, and the tracks can be regarded as the obstacles, which threatens the safety of the vehicle in the driving process.
Meanwhile, relevant trace enhancement boundary features are added according to needs, for example, trees on two sides of a road are a series of sparse points on a map, the area between the trees can cause influence when the area is locally planned, and the trees are mistaken to be passable paths, so that the boundaries can be manually processed, the sparse points are connected, and passable areas are corrected. Since the map is represented by a binary image, white represents an obstacle, and black represents a passable area. Covering the dynamic barrier part with black and artificially modifying the dynamic barrier part into a passable area; and connecting sparse points belonging to the obstacles by using a white part to enhance boundary constraint. The image editing software can select common editing software such as photoshop and the like, and finally the editing is finished, and the output in the same format as the original map is ensured.
D. Map splitting
The edited global map needs to be split into sub-maps again for local planning, and the position of each sub-map in the spliced global map is saved during picture splicing, including: the rotation angle, the length and the width after rotation and the position information of the rotated sub-map in the new map are split according to the data, the position information of each sub-map in the total map and the length and width information can be used for recovering the position information of the sub-map before translation, and the posture of the sub-map before translation is recovered by reversible transformation according to the rotation angle and eliminating the influence of a rotation matrix. The number of the split sub-maps is consistent with that of the sub-maps before splicing, the length of the sub-maps is consistent with that and width of the original sub-maps, and the vehicle pixel indexes and the pose information in the split sub-maps are consistent with those in the original sub-maps.
E. Calling mode
The sub-maps which are edited and split are sent out by the sensing module and are called as prior maps or offline maps, then the sub-maps are received and used by the sensing module, the processed maps are used for optimizing local path planning, and vehicles can plan by the prior information by the prior maps so as to make a plurality of prior decisions, so that the new performance of the system is improved.
Drawings
The attached figure 1 in the specification is an example of a splicing front part molecular map, the sub-maps store barrier information and passable areas, are local maps to be processed, and all the sub-maps need to be spliced and edited subsequently.
The attached figure 2 in the specification is a global map formed by splicing all local sub-maps according to pose information, barrier information and passable areas of a global environment are stored, and then the barrier information and the passable areas need to be edited to remove interference information and then are split.
Detailed Description
The following describes in detail preferred embodiments of the invention, in order to illustrate specific procedures for the use of the invention.
The invention discloses a method for acquiring a passable area based on a laser radar and editing and calling the passable area, which comprises the following steps of:
step S1, extracting and storing passable area
Acquiring and storing a passable area sub-map through a laser radar, and extracting and storing the total number of the sub-maps, the initial pose information of the sub-maps, the number of the current sub-maps, the resolution of the sub-maps, the height and width of pictures, the poses of the sub-maps, the pixel coordinates of vehicles in the sub-maps and the pose information of the pixel coordinates in a world coordinate system, which are output in the passable area extracting and storing process
Step S2, splicing sub-maps
And according to the saved sub-maps and the saved pose information data, modifying a path for storing the pose information file, the total number of the environment maps, the sequence numbers of the first sub-map and the last sub-map and a saving path of the target map in an image splicing program. And after the modification is finished, compiling the running program, finally obtaining a spliced global map and a global map marked with vehicle position and attitude points under the path of the target file, and simultaneously generating a text file in which the relevant information of the sub-map required for map disassembly in the global map is stored.
Step S3, editing and splicing the finished map
And editing the spliced map by using image processing software, removing dynamic obstacles such as pedestrians, running dynamic vehicles and the like in the map, and connecting sparse points such as trees and the like at two sides of a road to form a clear boundary constraint.
Step S4, splitting the edited map
According to the position information of each sub-map in the global map generated when the maps are spliced, the total number of the sub-maps needing to be split is modified in a map splitting program, the paths of the split sub-maps are saved, the storage paths of the edited global map and the text paths for storing the position information of the sub-maps are modified, compiling operation is carried out after modification is finished, and finally the split sub-maps are generated under a target folder.
Step S5, checking the split sub map
1) And judging whether the format and the name of the split sub-map are the same as those of the original sub-map or not.
2) The recorded text information should be identical to the original text information.
3) And placing the split sub-map at a specified position for the local planning module to call.
Step S6, calling method
The processed map is sent out through the sensing module, the planning module receives the processed map, and the planning module plans in the revised and completed passable area or modifies the global planning.
Claims (7)
1. A storage, efficient editing and calling method for a passable area of an unmanned vehicle is characterized by comprising the following steps:
acquiring an obstacle map, establishing a sub-map at intervals when the map is established due to a large passable area, storing the sub-map locally in the form of pictures, and storing the position and pose information of a vehicle in each sub-map, the pixel position of the vehicle in the map, the size of the map and the resolution;
and splicing all the sub-maps by using the position and pose information of each sub-map vehicle, wherein the splicing result is a global map containing all sub-map features, and the position information of the sub-maps in the new map is stored for use in map splitting.
Editing the spliced new graph.
And splitting and calling the edited new graph again.
2. The method for storing, efficiently editing and recalling an area accessible by unmanned vehicles according to claim 1, the step of obtaining an obstacle map comprising:
1) accepting messages for a plurality of different maps and for a milemeter, comprising: and after receiving the messages, carrying out clock synchronization on the messages so as to ensure the simultaneity of the messages.
2) And performing expansion processing on the obtained map.
3) And updating Bayesian probability, performing fusion estimation on all obstacle information and historical information of the current frame to obtain a new probability map, converting the probability map into a grid map, and performing morphological closed operation processing on the obtained grid map.
3. The step of obtaining a map of obstacles according to claim 2, the obtained passable area comprising the following information: the method comprises the steps of obtaining the total number of sub-maps, the initial pose information of the sub-maps, the current sub-map number, the resolution of the sub-maps, the height and the width of the sub-maps, the pixel coordinates of a vehicle in the sub-maps and the pose information of the pixel coordinates in a world coordinate system.
4. The method for storing, efficiently editing and calling the trafficable area of the unmanned vehicle as claimed in claim 1, wherein the step of using the position and pose information of each sub-map vehicle to splice all sub-maps comprises the following features:
1) and splicing the sub-maps on the premise of the saved position and pose information of the vehicle in each sub-map.
2) Each sub map is established under the vehicle body coordinate system at the sub map establishing time, the splicing process takes the acquired first sub map and the pose information as the reference, the rest sub maps are transformed, and the rest sub maps containing the vehicle pose information are unified to the vehicle body coordinate system of the first sub map for splicing. In the splicing process, an overlapping area may exist between the sub-map subjected to rotational translation and the existing map, and the intersection of the obstacles in the sub-map to be spliced and the existing map is taken as the obstacle in the spliced map. Since the obstacle appears as a white pixel, it appears in the image processing, and if there is a white pixel in any one of the images on the pixel, the pixel of the new image also takes the white pixel.
3) Unification of the coordinate systems requires a rotation matrix and a translation matrix between the coordinate systems. For a two-dimensional image, a rotation matrix can be obtained by the rotation angle. After the two-dimensional image is rotated, the length and width of the image are changed, so that the position, angle and length and width information of the origin are required to be stored, and the position information of the sub-map in the spliced image is determined. The position information embodies the content of the translation matrix, the rotation angle embodies the content of the rotation matrix, and the sub-map is spliced and simultaneously stores the related information of the sub-map in a new map, and the method mainly comprises the following steps: rotation angle, length and width after rotation, and position information of the rotated sub-map in the new map.
5. The method for storing, efficiently editing and calling the passable area for the unmanned vehicle as claimed in claim 1, wherein the new map after splicing is edited, the features to be processed mainly include dynamic obstacles, because the historical frame data are continuously superimposed, so that moving objects are continuously detected at different positions, and thus a moving track is formed, and the tracks are all regarded as obstacles, which greatly affects the safe traveling of vehicles, such as pedestrians, dynamically traveling vehicles and other dynamic obstacles, and meanwhile, boundary constraints with unobvious features are strengthened, such as trees at two sides of a road are a series of discontinuous tracks in the map, the boundary constraints are not obvious enough, and non-passable areas such as open lawns.
6. The method for storing, efficiently editing and calling the accessible area of the unmanned vehicle as claimed in claim 1, wherein the splitting and calling the edited new graph again mainly includes the following features:
1) and splitting according to the position information of the saved subgraphs in the new graph during splicing.
2) The number of the split sub-maps should be consistent with the number of the original sub-maps, and the length of the sub-maps should be consistent with the length and width of the original sub-maps.
3) The vehicle pixel index and the pose information in the split sub-map are consistent with those in the original sub-map.
7. The method for storing, efficiently editing and calling the accessible area for the unmanned vehicle according to claim 1, editing the spliced new map, wherein the edited map is called a priori map, and is sent by the sensing module, the planning module receives the edited map, and the planning module plans in the accessible area which is modified completely or modifies the global plan.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110101983.8A CN114791936A (en) | 2021-01-26 | 2021-01-26 | Storage, efficient editing and calling method for passable area of unmanned vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110101983.8A CN114791936A (en) | 2021-01-26 | 2021-01-26 | Storage, efficient editing and calling method for passable area of unmanned vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114791936A true CN114791936A (en) | 2022-07-26 |
Family
ID=82459746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110101983.8A Pending CN114791936A (en) | 2021-01-26 | 2021-01-26 | Storage, efficient editing and calling method for passable area of unmanned vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114791936A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116088503A (en) * | 2022-12-16 | 2023-05-09 | 深圳市普渡科技有限公司 | Dynamic obstacle detection method and robot |
-
2021
- 2021-01-26 CN CN202110101983.8A patent/CN114791936A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116088503A (en) * | 2022-12-16 | 2023-05-09 | 深圳市普渡科技有限公司 | Dynamic obstacle detection method and robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108983781B (en) | Environment detection method in unmanned vehicle target search system | |
| CN111666921B (en) | Vehicle control method, apparatus, computer device, and computer-readable storage medium | |
| Senlet et al. | A framework for global vehicle localization using stereo images and satellite and road maps | |
| WO2019153245A1 (en) | Systems and methods for deep localization and segmentation with 3d semantic map | |
| Yeap et al. | Computing a representation of the local environment | |
| CN108764187A (en) | Extract method, apparatus, equipment, storage medium and the acquisition entity of lane line | |
| US10402978B1 (en) | Method for detecting pseudo-3D bounding box based on CNN capable of converting modes according to poses of objects using instance segmentation and device using the same | |
| CN112950645B (en) | Image semantic segmentation method based on multitask deep learning | |
| CN109086277A (en) | A kind of overlay region building ground drawing method, system, mobile terminal and storage medium | |
| CN110986945B (en) | Local navigation method and system based on semantic altitude map | |
| WO2023123837A1 (en) | Map generation method and apparatus, electronic device, and storage medium | |
| CN113110455A (en) | Multi-robot collaborative exploration method, device and system for unknown initial state | |
| CN115639823A (en) | Terrain sensing and movement control method and system for robot under rugged and undulating terrain | |
| US10445611B1 (en) | Method for detecting pseudo-3D bounding box to be used for military purpose, smart phone or virtual driving based-on CNN capable of converting modes according to conditions of objects and device using the same | |
| CN117576652A (en) | Road object identification method and device, storage medium and electronic equipment | |
| US7542589B2 (en) | Road position detection | |
| Wang et al. | 3D-LIDAR based branch estimation and intersection location for autonomous vehicles | |
| CN109115232A (en) | The method and apparatus of navigation | |
| CN114791936A (en) | Storage, efficient editing and calling method for passable area of unmanned vehicle | |
| CN114187357A (en) | High-precision map production method and device, electronic equipment and storage medium | |
| CN113435227B (en) | Map generation and vehicle positioning method, system, device and storage medium | |
| CN115409910A (en) | Semantic map construction method, visual positioning method and related equipment | |
| CN117419738A (en) | Based on visual view and D * Path planning method and system of Lite algorithm | |
| Liu et al. | Vision-IMU multi-sensor fusion semantic topological map based on RatSLAM | |
| CN116523970A (en) | Dynamic three-dimensional target tracking method and device based on secondary implicit matching |
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 |