CN110231035B - Climbing mobile robot path guiding method - Google Patents
Climbing mobile robot path guiding method Download PDFInfo
- Publication number
- CN110231035B CN110231035B CN201910569062.7A CN201910569062A CN110231035B CN 110231035 B CN110231035 B CN 110231035B CN 201910569062 A CN201910569062 A CN 201910569062A CN 110231035 B CN110231035 B CN 110231035B
- Authority
- CN
- China
- Prior art keywords
- mobile robot
- point cloud
- climbing mobile
- climbing
- cloud data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Manipulator (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention relates to a climbing mobile robot path guiding method, relates to the technical field of robots, and is used for solving the technical problem that the image data updating speed of a robot in the prior art is slow. According to the path guiding method for the climbing mobile robot, noise points can be filtered out by carrying out noise reduction treatment on the three-dimensional point cloud data obtained by the binocular stereo camera, so that the obtained three-dimensional point cloud data are accurate; and the data volume can be reduced and the calculated amount can be reduced by filtering the three-dimensional point cloud data obtained by the binocular stereo camera, so that the updating speed of the three-dimensional point cloud data is increased, and the moving speed of the climbing mobile robot is increased.
Description
Technical Field
The invention relates to the technical field of robots, in particular to a climbing mobile robot path guiding method.
Background
If the robot wants to realize autonomous movement, navigation and positioning are needed in the environment. Autonomous navigation, as a core technology, is a key to endow the robot with perception and action capabilities, and relates to the problems of positioning and path planning. The visual navigation positioning is to acquire images through a sensing device on the robot body and complete high-level decisions such as image recognition, path planning and the like through a computer. The climbing robot for high-altitude operation has a complex working environment, so that the obtained image data volume is very large, the calculated amount is huge, the data update of each step of the robot is very slow, and the moving speed of the robot is seriously influenced.
Disclosure of Invention
The invention provides a climbing mobile robot path guiding method, which is used for solving the technical problem that the image data updating speed of a robot in the prior art is slow.
The invention provides a climbing mobile robot path guiding method, which comprises the following steps:
s1: a binocular stereo camera on the climbing mobile robot obtains an environment image, the environment image is calculated to obtain three-dimensional point cloud data, and the unit of the three-dimensional point cloud data is mm;
s2: denoising and filtering the three-dimensional point cloud data, and detecting an object by using the denoised and filtered three-dimensional point cloud data to identify a target and an obstacle in the environment;
s3: and enabling the climbing mobile robot to walk and/or cross in a straight line according to the detected information.
In one embodiment, in step S2, objects and obstacles in the environment are identified by identifying the number of planes.
In one embodiment, in step S2, a plane is extracted based on a random sample consensus algorithm.
In one embodiment, step S3 includes the following sub-steps:
s31: if the number of the planes identified in the step S2 is one, taking the direction of the point cloud boundary of the plane as the direction of the linear walking of the robot, so that the climbing mobile robot can walk linearly;
s32: and judging whether the walking direction of the climbing mobile robot is consistent with the direction of the point cloud boundary of the plane, if not, performing direction rectification treatment on both the front foot and the rear foot of the climbing mobile robot.
In one embodiment, step S3 includes the following sub-steps:
s31: if the number of the planes identified in the step S2 is two, the climbing mobile robot crosses according to an included angle between the two planes, and the direction of the point cloud boundary of one plane is taken as the walking direction of the climbing mobile robot, so that the climbing mobile robot walks linearly;
s32: and judging whether the walking direction of the climbing mobile robot is consistent with the direction of the point cloud boundary of the plane, if not, performing direction rectification treatment on both the front foot and the rear foot of the climbing mobile robot.
In one embodiment, step S2 includes the following sub-steps:
s21: removing noise points in the three-dimensional point cloud data by adopting a straight-through filtering algorithm and a statistical filtering algorithm;
s22: filtering the three-dimensional point cloud data without the noise points by a voxel grid downsampling algorithm to obtain filtered three-dimensional point cloud data;
s23: the three-dimensional point cloud data obtained in step S22 is used to identify objects and obstacles in the environment.
In one embodiment, the binocular stereo camera is provided with a laser displacement sensor, and the laser displacement sensor is used for measuring the distance from the binocular stereo camera to the ground and adjusting the depth threshold of the camera.
In one embodiment, the laser displacement sensor is disposed at an intermediate position on the housing of the binocular stereo camera, and the direction of the laser emitted from the laser displacement sensor is parallel to the direction of the optical axis of the binocular stereo camera.
In one embodiment, the binocular stereo camera comprises a left camera and a right camera, and the optical axis of the left camera is at an angle of 50 ° to 60 ° to the ground.
In one embodiment, the three-dimensional point cloud data is obtained by using the optical center of the left camera as an origin and performing coordinate system transformation to form three-dimensional data with the feet of the climbing robot as the origin.
Compared with the prior art, the invention has the advantages that: noise points can be filtered out by carrying out noise reduction treatment on three-dimensional point cloud data obtained by a binocular stereo camera, so that the obtained three-dimensional point cloud data is more accurate; and the data volume can be reduced and the calculated amount can be reduced by filtering the three-dimensional point cloud data obtained by the binocular stereo camera, so that the updating speed of the three-dimensional point cloud data is increased, and the moving speed of the climbing mobile robot is increased.
Drawings
The invention will be described in more detail hereinafter on the basis of embodiments and with reference to the accompanying drawings.
Fig. 1 is a schematic perspective view of a climbing mobile robot in an embodiment of the present invention;
FIG. 2 is an image captured by a binocular camera in an embodiment of the present invention;
FIG. 3 is a three-dimensional point cloud data plot obtained from images captured by a binocular camera in an embodiment of the present invention;
FIG. 4 is a three-dimensional point cloud data map having a plane in accordance with an embodiment of the present invention;
FIG. 5 is a three-dimensional point cloud data map of straight lines extracted in the plane shown in FIG. 4;
FIG. 6 is a three-dimensional point cloud data map with two planes in an embodiment of the invention;
fig. 7 is a flow chart of a method for guiding a path of a climbing mobile robot in an embodiment of the present invention.
Reference numerals:
1-binocular stereo camera; 2-laser displacement sensor, 3-climbing mobile robot.
Detailed Description
The invention will be further explained with reference to the drawings.
As shown in fig. 1 and 7, according to a first aspect of the present invention, there is provided a climbing mobile robot path guiding method including the following steps.
Firstly, an environment image is obtained by climbing a binocular stereo camera 1 on a mobile robot 3, and the environment image is calculated, so that three-dimensional point cloud data is obtained, wherein the unit of the three-dimensional point cloud data is mm.
Specifically, as shown in fig. 1, the binocular stereo camera 1 includes a left camera and a right camera disposed in parallel on the head of the climbing mobile robot 3, the left camera and the right camera corresponding to two eyes of the climbing mobile robot 3. The binocular camera captures images of the environment as shown in fig. 2, and the captured images are processed by a computer to form three-dimensional point cloud data as shown in fig. 3. The three-dimensional point cloud data is formed by taking the optical center of the left camera as an origin and converting a coordinate system, and the three-dimensional data taking the feet of the climbing robot as the origin is formed.
The data acquired by the method contains a lot of noise, which affects image recognition, so that the three-dimensional data needs to be subjected to noise reduction processing.
And secondly, denoising and filtering the three-dimensional point cloud data, and detecting objects by using the denoised and filtered three-dimensional point cloud data so as to identify targets and obstacles in the environment.
First, the procedure of the noise reduction processing in the present invention will be described in detail.
A part of the noise is filtered by using a straight-through filtering algorithm. Specifically, the distance ranges in the x, y, and z directions in the data are set separately, and noise that is not in this range is filtered out.
Then, the outliers are removed by using a statistical filtering method. Specifically, the filtering idea is to perform a statistical analysis on the neighborhood of each point and calculate its average distance to all nearby points. Assuming that the result is a gaussian distribution whose shape is determined by the mean and standard deviation, then points whose mean distance is outside the standard range (defined by the global distance mean and variance) can be defined as outliers and removed from the data.
Further, although noise is removed, since the amount of data is very large, which affects the calculation speed, it is necessary to reduce the data by a voxel grid down-sampling algorithm, and the accuracy of the data is not greatly affected. Specifically, a three-dimensional voxel grid is created for the point cloud data from which noise points are removed (the voxel grid can be regarded as a set of tiny spatial three-dimensional cubes), then in each voxel (i.e. each three-dimensional cube), the center of gravity of all the points in the voxel is used for approximately displaying other points in the voxel, so that all the points in the voxel are represented by a center of gravity point, and the filtered three-dimensional point cloud data obtained after all the voxels are processed can be greatly reduced in computation amount through the filtering, so that the computation speed is increased.
It should be noted that the three-dimensional point cloud data is updated every time the climbing mobile robot 3 walks.
Secondly, the present invention identifies objects and obstacles in the environment by identifying the number of planes. The process of identifying planes in the present invention will be described in detail below.
As shown in fig. 4, a plane is identified by using the filtered three-dimensional point cloud data. The plane is extracted based on ANSAC (random sample consensus) algorithm. Specifically, n points are randomly selected from the filtered three-dimensional point cloud data and are marked as interior points; the remaining points are marked as outer points.
Fitting a plane by using the n interior points, and setting a distance threshold d from one point to the plane1If exterior pointThe distance to this plane being less than d1If so, marking the outer point as an inner point, and recording the distance error from the outer point to the plane; and randomly repeating the process, and iterating for k times, selecting the plane with the minimum error, wherein the number of inner points in the plane with the minimum error is more than a certain number m.
Further, the number of iterations k is determined. If p represents the probability that all the randomly selected points in the data set are inner points in the iteration process, 1-p represents the probability that at least one randomly selected point in the data set is an outer point in the iteration process; if q represents the probability of selecting an interior point from the dataset each time, assuming that the estimation model requires the selection of n points, qnIs the probability that all n points are interior points; then 1-qnIs the probability that at least one of the n points is an outlier, indicating that a poor model was estimated from the data set.
(1-qn) k represents the probability that the algorithm never chooses "n points are interior points", which is the same as 1-p, i.e.:
(1-qn)k＝1-p；
taking logarithm of two sides of the above expression, it can be known that k satisfies the following expression:
and providing information for the next action of climbing the mobile robot 3 according to the number of the extracted planes.
And thirdly, enabling the climbing mobile robot 3 to walk linearly and/or cross according to the detected information.
In one embodiment, if the number of the planes identified in the previous step is one, the climbing mobile robot 3 is caused to travel linearly by taking the direction of the point cloud boundary of the plane as the direction in which the robot travels linearly. As shown in fig. 4, there is a plane in the three-dimensional data point cloud, and if a plane is identified, the climbing mobile robot 3 walks in a straight line.
The walking direction of the climbing mobile robot 3 is determined according to two side lines of the plane, and the position of the climbing mobile robot 3 is adjusted according to the direction. As shown in fig. 5, a red dot (a dot shown as a darker color in the black-and-white line graph) is the plane boundary point.
Specifically, a method of extracting a straight line from the plane is as follows. The three-dimensional point cloud data is represented by a kd-tree (k-dimensional tree), the normal features of the three-dimensional point cloud data are calculated, then boundary points are estimated, and then a straight line is extracted according to the principle of a random sampling consistency algorithm and is used as the next walking direction of the climbing mobile robot 3.
Further, whether the walking direction of the climbing mobile robot 3 is consistent with the direction of the point cloud boundary of the plane or not is judged, if not, the direction correction processing is carried out on the front foot and the rear foot of the climbing mobile robot 3, and if yes, the walking can be continued.
In one embodiment, if the number of the planes identified in the previous step is two, the climbing mobile robot 3 crosses according to an included angle between the two planes, and the direction of the point cloud boundary of one of the planes is taken as the direction in which the climbing mobile robot 3 travels, so that the climbing mobile robot 3 travels linearly. As shown in fig. 6, the three-dimensional point cloud data includes two planes, and if the two planes are identified, the angles of the two planes and the distance between the origin of the climbing mobile robot 3 and the intersection line of the two planes are calculated, so that information is provided for the robot to walk across, and the robot adopts different crossing strategies.
Further, whether the walking direction of the climbing mobile robot 3 is consistent with the direction of the point cloud boundary of the plane or not is judged, if not, the direction correction processing is carried out on the front foot and the rear foot of the climbing mobile robot 3, and if yes, the walking can be continued.
In one embodiment, the binocular stereo camera 1 is provided with the laser displacement sensor 2, the laser displacement sensor 2 can measure the distance between the binocular stereo camera 1 and the ground, and the positions of the binocular stereo camera 1 and the laser displacement sensor 2 are determined, so that the distance between the binocular stereo camera 1 and the ground can be obtained, the distance threshold value when the binocular stereo camera 1 forms a three-dimensional point cloud is adjusted, the binocular stereo camera 1 is enabled to be in a working state, and the working efficiency is improved.
Specifically, the laser displacement sensor 2 is disposed at an intermediate position on the housing of the binocular stereo camera 1, as shown in fig. 1, and the direction of the laser light emitted by the laser displacement sensor 1 is parallel to the optical axis direction of the binocular stereo camera 1.
Further, the optical axis of the left camera is at an angle of 50-60, for example 55, to the ground.
Because the binocular stereo camera 1 needs to be installed at a proper height and angle to obtain a proper view field, the angle of the binocular stereo camera 1 can be conveniently adjusted through the distance data collected by the laser displacement sensor 2, and the depth range of the binocular stereo camera 1 is measured.
In one embodiment, the baseline of the left and right cameras is 100mm, the focal length is 3.6mm, and the resolution is 752 × 480.
In conclusion, the climbing mobile robot path guiding method can realize real-time positioning, position correction and path planning of the climbing robot, and has the characteristics of high precision, high calculation speed, high reliability and the like. The binocular stereo camera 1 provides visual information for walking, climbing and crossing of the robot, and can greatly reduce calculated amount through noise reduction and filtering processing, so that extracted useful information is sent to a control system of the climbing mobile robot, objects are identified according to the received information, and real-time positioning and path planning are achieved.
While the invention has been described with reference to a preferred embodiment, various modifications may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In particular, the technical features mentioned in the embodiments can be combined in any way as long as there is no structural conflict. It is intended that the invention not be limited to the particular embodiments disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.
Claims (9)
1. A climbing mobile robot path guiding method is characterized by comprising the following steps:
s1: obtaining an environment image through a binocular stereo camera on a climbing mobile robot, and calculating the environment image to obtain three-dimensional point cloud data, wherein the unit of the three-dimensional point cloud data is mm;
s2: denoising and filtering the three-dimensional point cloud data, and detecting an object by using the denoised and filtered three-dimensional point cloud data to identify a target and an obstacle in the environment;
s3: enabling the climbing mobile robot to walk linearly and/or cross according to the detected information;
step S3 includes the following substeps:
s31: if the number of the planes identified in the step S2 is one, taking the direction of the point cloud boundary of the plane as the direction of the linear walking of the robot, so that the climbing mobile robot can walk linearly;
s32: and judging whether the walking direction of the climbing mobile robot is consistent with the direction of the point cloud boundary of the plane, if not, performing direction rectification treatment on both the front foot and the rear foot of the climbing mobile robot.
2. The climbing mobile robot path guiding method according to claim 1, wherein in step S2, objects and obstacles in the environment are identified by identifying the number of planes.
3. The path guidance method for the climbing mobile robot according to claim 2, wherein in step S2, the plane is extracted based on a random sampling consensus algorithm.
4. The climbing mobile robot path guiding method according to any one of claims 1 to 3, wherein step S3 comprises the sub-steps of:
s31: if the number of the planes identified in the step S2 is two, the climbing mobile robot crosses according to an included angle between the two planes, and the direction of the point cloud boundary of one plane is taken as the walking direction of the climbing mobile robot, so that the climbing mobile robot walks linearly;
s32: and judging whether the walking direction of the climbing mobile robot is consistent with the direction of the point cloud boundary of the plane, if not, performing direction rectification treatment on both the front foot and the rear foot of the climbing mobile robot.
5. The climbing mobile robot path guiding method according to any one of claims 1 to 3, wherein step S2 comprises the sub-steps of:
s21: removing noise points in the three-dimensional point cloud data by adopting a straight-through filtering algorithm and a statistical filtering algorithm;
s22: filtering the three-dimensional point cloud data without the noise points by a voxel grid downsampling algorithm to obtain filtered three-dimensional point cloud data;
s23: the three-dimensional point cloud data obtained in step S22 is used to identify objects and obstacles in the environment.
6. The climbing mobile robot path guiding method according to any one of claims 1 to 3, wherein a laser displacement sensor is arranged on the binocular stereo camera, and the laser displacement sensor is used for measuring the distance from the binocular stereo camera to the ground and adjusting a depth threshold of the binocular stereo camera.
7. The climbing mobile robot path guiding method according to claim 6, wherein the laser displacement sensor is disposed at an intermediate position on the housing of the binocular stereo camera, and a direction of the laser light emitted by the laser displacement sensor is parallel to an optical axis direction of the binocular stereo camera.
8. The climbing mobile robot path guiding method according to claim 7, wherein the binocular stereo camera comprises a left camera and a right camera, and an angle between an optical axis of the left camera and the ground is 50-60 °.
9. The method as claimed in claim 8, wherein the three-dimensional point cloud data is transformed from a coordinate system with the optical center of the left camera as an origin to form three-dimensional data with the foot of the climbing robot as the origin.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910569062.7A CN110231035B (en) | 2019-06-27 | 2019-06-27 | Climbing mobile robot path guiding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910569062.7A CN110231035B (en) | 2019-06-27 | 2019-06-27 | Climbing mobile robot path guiding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110231035A CN110231035A (en) | 2019-09-13 |
| CN110231035B true CN110231035B (en) | 2020-03-20 |
Family
ID=67857482
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910569062.7A Active CN110231035B (en) | 2019-06-27 | 2019-06-27 | Climbing mobile robot path guiding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110231035B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113867333B (en) * | 2021-09-03 | 2023-11-17 | 南方科技大学 | Four-foot robot stair climbing planning method based on visual perception and application thereof |
| CN113917917B (en) * | 2021-09-24 | 2023-09-15 | 四川启睿克科技有限公司 | Obstacle avoidance method and device for indoor bionic multi-legged robot and computer readable medium |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107179768B (en) * | 2017-05-15 | 2020-01-17 | 上海木木机器人技术有限公司 | Obstacle identification method and device |
| CN107656545A (en) * | 2017-09-12 | 2018-02-02 | 武汉大学 | A kind of automatic obstacle avoiding searched and rescued towards unmanned plane field and air navigation aid |
| CN108984741B (en) * | 2018-07-16 | 2021-06-04 | 北京三快在线科技有限公司 | Map generation method and device, robot and computer-readable storage medium |
| CN108858199B (en) * | 2018-07-27 | 2020-04-07 | 中国科学院自动化研究所 | Method for grabbing target object by service robot based on vision |
| CN109141364B (en) * | 2018-08-01 | 2020-11-03 | 北京进化者机器人科技有限公司 | Obstacle detection method and system and robot |
| CN109102539A (en) * | 2018-08-15 | 2018-12-28 | 宝钢湛江钢铁有限公司 | A kind of stock yard based on 3 D laser scanning automatically generates stock ground drawing method |
| CN109190704A (en) * | 2018-09-06 | 2019-01-11 | 中国科学院深圳先进技术研究院 | The method and robot of detection of obstacles |
| CN109144072A (en) * | 2018-09-30 | 2019-01-04 | 亿嘉和科技股份有限公司 | A kind of intelligent robot barrier-avoiding method based on three-dimensional laser |
| CN109872324A (en) * | 2019-03-20 | 2019-06-11 | 苏州博众机器人有限公司 | Ground obstacle detection method, device, equipment and storage medium |
| CN109857123A (en) * | 2019-03-21 | 2019-06-07 | 郑州大学 | A kind of fusion method of view-based access control model perception and the indoor SLAM map of laser acquisition |
-
2019
- 2019-06-27 CN CN201910569062.7A patent/CN110231035B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110231035A (en) | 2019-09-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fankhauser et al. | Probabilistic terrain mapping for mobile robots with uncertain localization | |
| US10275649B2 (en) | Apparatus of recognizing position of mobile robot using direct tracking and method thereof | |
| US10399228B2 (en) | Apparatus for recognizing position of mobile robot using edge based refinement and method thereof | |
| CN105021124B (en) | A kind of planar part three-dimensional position and normal vector computational methods based on depth map | |
| CN112767490B (en) | Outdoor three-dimensional synchronous positioning and mapping method based on laser radar | |
| CN109857123A (en) | A kind of fusion method of view-based access control model perception and the indoor SLAM map of laser acquisition | |
| EP3159122A1 (en) | Device and method for recognizing location of mobile robot by means of search-based correlation matching | |
| CN110231035B (en) | Climbing mobile robot path guiding method | |
| CN113804184A (en) | Ground robot positioning method based on multiple sensors | |
| CN113592891B (en) | Unmanned vehicle passable domain analysis method and navigation grid map manufacturing method | |
| CN115592324A (en) | Automatic welding robot control system based on artificial intelligence | |
| CN115451948A (en) | Agricultural unmanned vehicle positioning odometer method and system based on multi-sensor fusion | |
| CN114066773B (en) | Dynamic object removal based on point cloud characteristics and Monte Carlo expansion method | |
| CN109544632B (en) | Semantic SLAM object association method based on hierarchical topic model | |
| CN113487631B (en) | LEGO-LOAM-based adjustable large-angle detection sensing and control method | |
| CN110851978A (en) | Camera position optimization method based on visibility | |
| CN114004900A (en) | Indoor binocular vision odometer method based on point-line-surface characteristics | |
| JP2016177742A (en) | Mobile body control device, landmark, and program | |
| CN117075158A (en) | Pose estimation method and system of unmanned deformation motion platform based on laser radar | |
| Wang et al. | Micro aerial vehicle navigation with visual-inertial integration aided by structured light | |
| Souza et al. | Probabilistic robotic grid mapping based on occupancy and elevation information | |
| CN114648571A (en) | Method for filtering obstacles in driving area in high-precision mapping of robot | |
| CN114066824B (en) | Binocular vision odometer method with dynamic target detection function | |
| CN117173376B (en) | Mobile track planning method and system for medical equipment | |
| CN112148032B (en) | Miniature tree hole detection aircraft and control method |
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 |