CN110763223A - Sliding window based indoor three-dimensional grid map feature point extraction method - Google Patents
Sliding window based indoor three-dimensional grid map feature point extraction method Download PDFInfo
- Publication number
- CN110763223A CN110763223A CN201911049501.8A CN201911049501A CN110763223A CN 110763223 A CN110763223 A CN 110763223A CN 201911049501 A CN201911049501 A CN 201911049501A CN 110763223 A CN110763223 A CN 110763223A
- Authority
- CN
- China
- Prior art keywords
- sliding window
- grid
- dimensional grid
- grid map
- window
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
- G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a sliding window-based indoor three-dimensional grid map feature point extraction method, which comprises the following steps: step 1, obtaining a three-dimensional grid map of an indoor environment by using a graph-SLAM algorithm; step 2, establishing a sliding window and calculating a duty ratio; step 3, moving the sliding window in the step 2 in any direction, and if the duty ratio of the sliding window is obviously changed, determining the sliding window as a primary characteristic point; step 4, carrying out plane segmentation on the three-dimensional grid map in the step 1 by using a method based on region growing; step 5, performing plane fitting on the three-dimensional grid segmented in the step 4; step 6, calculating the distances from the preliminary feature points in the step 3 to three planes adjacent to the preliminary feature points, and if the distances are smaller than a threshold value, determining the distances as correct feature points; and 7, clustering the correct characteristic points in the step 6. The method can solve the problems of front-end construction based on graph optimization SLAM and multi-robot SLAM map fusion.
Description
Technical Field
The invention belongs to the technical field of mobile robot map construction, in particular to a single-robot graph-SLAM and multi-robot SLAM technology, and particularly relates to a sliding window-based indoor three-dimensional grid map feature point extraction method.
Background
When the mobile robot works in an indoor environment and cannot obtain pose information through external equipment (such as a GPS), at the moment, the robot can sense the surrounding environment through a sensor carried by the robot and create a map, and then the robot is positioned according to the created map, namely, the mobile robot is positioned and mapped (SLAM) at the same time, and on the basis, the mobile robot can complete other works, such as exploration, path planning, navigation and the like. Therefore, SLAM is the basis for the mobile robot to perform complex tasks in an unknown environment, and is the key to the intelligence of the mobile robot.
The SLAM algorithm of the mobile robot is mainly classified into a filtering-based method and a graph-based optimization method. The filtering-based method can only predict and update the current state of the robot, if the state of the robot has an error at a certain moment, the error will accompany the whole process of the mobile robot to construct the image, so the filtering-based method is not suitable for large-scale environment mapping. Unlike the traditional filtering-based method, graph-SLAM uses an optimization algorithm to realize SLAM after collecting the information of the complete map. The graph-SLAM algorithm can be divided into a front end and a back end, wherein the front end is responsible for the construction of the graph and mainly comprises two processes of sequential data association and loop closed-loop detection. The front-end constructed graph is called a pose graph (position-graph), the rear end of the graph-SLAM is responsible for using an optimizer to carry out global optimization on the pose graph, and closed-loop detection is carried out on the optimized pose graph until the pose graph is not updated any more.
The map can be divided into a two-dimensional map and a three-dimensional map, the three-dimensional map information is richer, and the environmental information can be truly embodied. At present, three-dimensional maps mainly include a point cloud map and a three-dimensional grid map. The point cloud map can only reflect the surface information of an object, and the environment is not divided into an occupied area, an idle area and an unknown area, so that the point cloud map cannot be used for navigation of the mobile robot, the three-dimensional grid map can accurately reflect real environment information, and map information can be provided for the mobile robot to complete complex tasks.
Therefore, the method for researching the characteristic point problem of the three-dimensional grid map has important significance for the fusion of the single-robot and multi-robot three-dimensional grid maps.
Disclosure of Invention
The invention aims to: the method is used for solving the problems of front-end construction based on graph optimization SLAM and multi-robot SLAM map fusion.
The technical scheme of the invention is as follows: a sliding window based indoor three-dimensional grid map feature point extraction method comprises the following steps:
step 1, obtaining a three-dimensional grid map of an indoor environment by using a graph-SLAM algorithm;
step 2, establishing a sliding window and calculating a duty ratio;
step 3, moving the sliding window in the step 2 in any direction, and if the duty ratio of the sliding window is obviously changed, determining the sliding window as a primary characteristic point;
step 4, carrying out plane segmentation on the three-dimensional grid map in the step 1 by using a method based on region growing;
step 5, performing plane fitting on the three-dimensional grid segmented in the step 4;
step 6, calculating the distances from the preliminary feature points in the step 3 to three planes adjacent to the preliminary feature points, and if the distances are smaller than a threshold value, determining the distances as correct feature points;
and 7, clustering the correct characteristic points in the step 6.
In the above, the indoor environment may be selected from a room, a corridor, and the like.
In the technical scheme, in the step 1, the value range of each grid of the three-dimensional grid map is 0-1. Where closer to 0 indicates that the grid is more likely to be free, closer to 1 indicates that the grid is more likely to occupy, and 0.5 indicates an unknown region.
In the above technical solution, in step 2, a sliding window of 7 × 7 is established with each solid grid as a center.
In the above technical solution, the distribution of the stereoscopic grids in the sliding window is analyzed, and the ratio of the number of occupied grids in the window to the number of idle grids is calculated, wherein the occupied grids are defined as 1, and the idle grids are defined as 0.
In the above technical solution, in step 4, the method based on region growing is to start from a grid with the smallest curvature, and determine whether a neighborhood grid and the grid belong to a plane according to the characteristics of the neighborhood grid.
In the above technical solution, the characteristics of the domain grid include a normal direction and a curvature value.
In the above technical solution, in step 6, the distance is constrained to be:
wherein A, B, C, D is a parameter of the plane equation, x0、y0、z0And e is the coordinate of the three-dimensional grid, wherein the epsilon is a distance threshold value, and when the distance threshold values from the initial characteristic point to the three adjacent planes are smaller than the epsilon, the grid is a correct characteristic point.
In the above technical solution, in step 7, the extracted feature points are clustered, and only a small number of feature points are used for characterization at the same corner point.
The feature points of the present invention are defined as points where three planes intersect, such as corners, and are used to characterize a three-dimensional volumetric grid map.
The invention has the advantages that:
the method comprises the steps of extracting characteristic points from an indoor three-dimensional grid map by using a sliding window-based method, extracting preliminary characteristic points according to different duty ratios of internal angle points, plane points and edge points of a window, then eliminating wrong characteristic points according to the distances from the preliminary characteristic points to three adjacent faces, and finally clustering the extracted characteristic points, wherein only a small number of characteristic points are used for representing the same angle point, so that the problems of front-end construction based on graph optimization SLAM and multi-robot SLAM map fusion are solved.
Drawings
The invention is further described with reference to the following figures and examples:
FIG. 1 is a flowchart of a method according to a first embodiment of the present invention
Fig. 2 is a three-dimensional grid map according to a first embodiment of the invention.
Fig. 3 is a schematic diagram illustrating the definition of feature points according to the first embodiment of the present invention.
Fig. 4 is a schematic diagram of extracting preliminary feature points of a local map based on a sliding window according to a first embodiment of the present invention.
Fig. 5 is a schematic diagram of feature points extracted by a sliding window-based stereo grid according to a first embodiment of the present invention.
Fig. 6 is a schematic diagram of feature point clustering according to a first embodiment of the present invention.
Detailed Description
The first embodiment is as follows:
referring to fig. 1, the invention relates to a sliding window-based method for extracting feature points of an indoor three-dimensional grid map, which comprises the following steps:
step 1, obtaining a three-dimensional grid map of an indoor environment by using a graph-SLAM algorithm;
referring to fig. 2, the three-dimensional grid map is obtained according to laser sensor and mobile robot IMU data;
referring to fig. 3, the definition of the feature points of the present invention is: the points where the three surfaces intersect, such as corners, are used to characterize the three-dimensional volumetric grid map.
Step 2, establishing a sliding window and calculating a duty ratio;
and establishing a 7 × 7 sliding window by taking each three-dimensional grid as a center, analyzing the distribution of the three-dimensional grids in the sliding window, and calculating the ratio of the number of occupied grids in the window to the number of idle grids, wherein the occupied grids are defined as 1, and the idle grids are defined as 0.
Step 3, moving the sliding window in the step 2 in any direction, judging that the grid is a plane grid, an edge grid or a feature point grid according to the change of the duty ratio in the sliding window, and if the duty ratios of the grids are obviously changed, determining that the grid is a primary feature point;
referring to fig. 4, a red grid (i.e., a grid corresponding to a color) in the graph indicates that, using preliminary feature points extracted based on a sliding window, it can be seen from the graph that feature points can be extracted at corner points, but there are wrong feature points.
Step 4, carrying out plane segmentation on the three-dimensional grid map in the step 1 by using a method based on region growing;
the method based on region growing is to start from a grid with the minimum curvature, and judge whether a neighborhood grid and the grid belong to a plane according to the characteristics (normal direction and curvature value) of the neighborhood grid.
Step 5, performing plane fitting on the three-dimensional grid segmented in the step 4;
step 6, calculating the distances from the preliminary feature points in the step 3 to three planes adjacent to the preliminary feature points, and if the distances are smaller than a threshold value, determining the distances as correct feature points;
in particular, the distance is constrained to be:
wherein A, B, C, D is a parameter of the plane equation, x0、y0、z0And e is the coordinate of the three-dimensional grid, wherein the epsilon is a distance threshold value, and when the distance threshold values from the initial characteristic point to the three adjacent planes are smaller than the epsilon, the grid is a correct characteristic point.
Referring to fig. 5, a red grid (i.e., a grid of a color corresponding to B) in the graph represents a correct feature point after an erroneous feature point is removed, and it can be seen from the graph that many feature points are distributed at the same corner, which may cause repeated calculation of feature point matching and affect algorithm efficiency.
And 7, clustering the correct feature points in the step 6, and representing the feature points at the same corner point by using a small number of feature points.
Referring to fig. 6, a red grid (i.e., a grid of C corresponding to color) in the graph represents the clustered feature points. Clustering the characteristic points with similar distances, and replacing the previous characteristic points with the geometric centers of the characteristic points to reduce the calculation amount of the following algorithm.
The indoor environment can be selected from scenes such as rooms, corridors and the like.
In this embodiment, the numeric area of each grid of the three-dimensional grid map in step 1 is 0-1. Where closer to 0 indicates that the grid is more likely to be free, closer to 1 indicates that the grid is more likely to occupy, and 0.5 indicates an unknown region.
It should be understood that the above-mentioned embodiments are only illustrative of the technical concepts and features of the present invention, and are intended to enable those skilled in the art to understand the contents of the present invention and implement the present invention, and not to limit the scope of the present invention. All modifications made according to the spirit of the main technical scheme of the invention are covered in the protection scope of the invention.
Claims (7)
1. A sliding window based indoor three-dimensional grid map feature point extraction method is characterized by comprising the following steps:
step 1, obtaining a three-dimensional grid map of an indoor environment by using a graph-SLAM algorithm;
step 2, establishing a sliding window and calculating a duty ratio;
step 3, moving the sliding window in the step 2 in any direction, and if the duty ratio of the sliding window is obviously changed, determining the sliding window as a primary characteristic point;
step 4, carrying out plane segmentation on the three-dimensional grid map in the step 1 by using a method based on region growing;
step 5, performing plane fitting on the three-dimensional grid segmented in the step 4;
step 6, calculating the distances from the preliminary feature points in the step 3 to three planes adjacent to the preliminary feature points, and if the distances are smaller than a threshold value, determining the distances as correct feature points;
and 7, clustering the correct characteristic points in the step 6.
2. The sliding-window-based indoor three-dimensional grid map feature point extraction method according to claim 1, wherein in step 1, a range of values of each grid of the three-dimensional grid map is 0-1.
3. The sliding-window-based indoor three-dimensional grid map feature point extraction method according to claim 1, wherein in step 2, a sliding window of 7 × 7 is established with each grid as a center.
4. The sliding-window-based indoor three-dimensional stereoscopic grid map feature point extraction method of claim 3, wherein the distribution of the stereoscopic grids in the sliding window is analyzed, and the ratio of the number of occupied grids in the window to the number of idle grids is calculated, wherein the occupied grids are defined as 1, and the idle grids are defined as 0.
5. The method for extracting feature points of an indoor three-dimensional grid map based on a sliding window according to claim 1, wherein in step 4, the method based on region growing is to determine whether a neighborhood grid and a grid thereof belong to a plane according to characteristics of the neighborhood grid starting from the grid with the smallest curvature.
6. The sliding-window-based indoor three-dimensional stereoscopic grid map feature point extraction method of claim 5, wherein the characteristics of the domain grid comprise a normal direction and a curvature value.
7. The sliding-window-based indoor three-dimensional grid map feature point extraction method according to claim 1, wherein in step 6, the distance is constrained to be:
wherein A, B, C, D is a parameter of the plane equation, x0、y0、z0And e is the coordinate of the three-dimensional grid, wherein the epsilon is a distance threshold value, and when the distance threshold values from the initial characteristic point to the three adjacent planes are smaller than the epsilon, the grid is a correct characteristic point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911049501.8A CN110763223B (en) | 2019-10-31 | 2019-10-31 | Sliding window based indoor three-dimensional grid map feature point extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911049501.8A CN110763223B (en) | 2019-10-31 | 2019-10-31 | Sliding window based indoor three-dimensional grid map feature point extraction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110763223A true CN110763223A (en) | 2020-02-07 |
| CN110763223B CN110763223B (en) | 2022-03-18 |
Family
ID=69334964
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911049501.8A Active CN110763223B (en) | 2019-10-31 | 2019-10-31 | Sliding window based indoor three-dimensional grid map feature point extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110763223B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273645A (en) * | 2022-08-09 | 2022-11-01 | 南京大学 | Map making method for automatically clustering indoor surface elements |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105184243A (en) * | 2015-08-24 | 2015-12-23 | 王红军 | Environment characteristic expression and identification method based on three dimensional grid map |
| CN107092020A (en) * | 2017-04-19 | 2017-08-25 | 北京大学 | Merge the surface evenness monitoring method of unmanned plane LiDAR and high score image |
| CN107316003A (en) * | 2017-06-06 | 2017-11-03 | 银江股份有限公司 | A kind of automobile logo detection method and system based on independent variable linear regression and structure tensor determinant |
| CN108416377A (en) * | 2018-02-26 | 2018-08-17 | 阿博茨德（北京）科技有限公司 | Information extracting method in block diagram and device |
| CN109330501A (en) * | 2018-11-30 | 2019-02-15 | 深圳乐动机器人有限公司 | A kind of method and sweeping robot cleaning ground |
| US20190095764A1 (en) * | 2017-09-26 | 2019-03-28 | Panton, Inc. | Method and system for determining objects depicted in images |
| CN110110763A (en) * | 2019-04-19 | 2019-08-09 | 苏州大学 | A kind of grating map fusion method based on maximum public subgraph |
| CN110390664A (en) * | 2018-11-30 | 2019-10-29 | 武汉滨湖电子有限责任公司 | One kind being based on the recognition methods of holes filling pavement crack |
-
2019
- 2019-10-31 CN CN201911049501.8A patent/CN110763223B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105184243A (en) * | 2015-08-24 | 2015-12-23 | 王红军 | Environment characteristic expression and identification method based on three dimensional grid map |
| CN107092020A (en) * | 2017-04-19 | 2017-08-25 | 北京大学 | Merge the surface evenness monitoring method of unmanned plane LiDAR and high score image |
| CN107316003A (en) * | 2017-06-06 | 2017-11-03 | 银江股份有限公司 | A kind of automobile logo detection method and system based on independent variable linear regression and structure tensor determinant |
| US20190095764A1 (en) * | 2017-09-26 | 2019-03-28 | Panton, Inc. | Method and system for determining objects depicted in images |
| CN108416377A (en) * | 2018-02-26 | 2018-08-17 | 阿博茨德（北京）科技有限公司 | Information extracting method in block diagram and device |
| CN109330501A (en) * | 2018-11-30 | 2019-02-15 | 深圳乐动机器人有限公司 | A kind of method and sweeping robot cleaning ground |
| CN110390664A (en) * | 2018-11-30 | 2019-10-29 | 武汉滨湖电子有限责任公司 | One kind being based on the recognition methods of holes filling pavement crack |
| CN110110763A (en) * | 2019-04-19 | 2019-08-09 | 苏州大学 | A kind of grating map fusion method based on maximum public subgraph |
Non-Patent Citations (2)
| Title |
|---|
| ANH-VUVO等: "Octree-based region growing for point cloud segmentation", 《ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING》 * |
| YU ZHONG: "Intrinsic shape signatures: a shape descriptor for 3D object recognition", 《2009 IEEE 12TH INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS, ICCV WORKSHOPS》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273645A (en) * | 2022-08-09 | 2022-11-01 | 南京大学 | Map making method for automatically clustering indoor surface elements |
| CN115273645B (en) * | 2022-08-09 | 2024-04-09 | 南京大学 | Map making method for automatically clustering indoor surface elements |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110763223B (en) | 2022-03-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111563442B (en) | Slam method and system for fusing point cloud and camera image data based on laser radar | |
| US11816907B2 (en) | Systems and methods for extracting information about objects from scene information | |
| CN111429563B (en) | Pipeline three-dimensional reconstruction method, system, medium and equipment based on deep learning | |
| CN113593017B (en) | Method, device, equipment and storage medium for constructing surface three-dimensional model of strip mine | |
| CN109949326A (en) | Contour of building line drawing method based on Backpack type three-dimensional laser point cloud data | |
| CN108133496B (en) | Dense map creation method based on g2o and random fern algorithm | |
| CN109035329A (en) | Camera Attitude estimation optimization method based on depth characteristic | |
| CN109960269A (en) | A kind of simplification environmental modeling method for serving unmanned plane independent navigation | |
| Quintana et al. | Door detection in 3D colored laser scans for autonomous indoor navigation | |
| CN114677323A (en) | Semantic vision SLAM positioning method based on target detection in indoor dynamic scene | |
| CN114279433A (en) | Map data automatic production method, related device and computer program product | |
| CN110967020B (en) | Simultaneous drawing and positioning method for port automatic driving | |
| CN114186112B (en) | Robot navigation method based on Bayesian optimization multiple information gain exploration strategy | |
| CN113902862B (en) | Visual SLAM loop verification system based on consistency cluster | |
| CN110763223B (en) | Sliding window based indoor three-dimensional grid map feature point extraction method | |
| CN113689459B (en) | Real-time tracking and mapping method based on GMM and YOLO under dynamic environment | |
| Wang et al. | Multidimensional particle swarm optimization-based unsupervised planar segmentation algorithm of unorganized point clouds | |
| CN113296504A (en) | Mobile robot mapping and path planning method based on RGBD depth camera | |
| CN115454055B (en) | Multi-layer fusion map representation method for indoor autonomous navigation and operation | |
| Zhang et al. | 3D reconstruction of weak feature indoor scenes based on hector SLAM and floorplan generation | |
| Zhang et al. | An improvement algorithm for OctoMap based on RGB-D SLAM | |
| Hochdorfer et al. | Landmark rating and selection according to localization coverage: Addressing the challenge of lifelong operation of SLAM in service robots | |
| Qiu et al. | A sliding window based feature extraction method with its application to 3D grid map fusion | |
| Wu et al. | OC-SLAM: steadily tracking and mapping in dynamic environments | |
| CN117537803B (en) | Robot inspection semantic-topological map construction method, system, equipment and 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 |