CN112747749B - Positioning navigation system based on binocular vision and laser fusion - Google Patents
Positioning navigation system based on binocular vision and laser fusion Download PDFInfo
- Publication number
- CN112747749B CN112747749B CN202011537391.2A CN202011537391A CN112747749B CN 112747749 B CN112747749 B CN 112747749B CN 202011537391 A CN202011537391 A CN 202011537391A CN 112747749 B CN112747749 B CN 112747749B
- Authority
- CN
- China
- Prior art keywords
- pose
- laser
- module
- sub
- matching
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 57
- 230000000007 visual effect Effects 0.000 claims abstract description 42
- 238000013507 mapping Methods 0.000 claims abstract description 15
- 238000013075 data extraction Methods 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims abstract description 9
- 238000000034 method Methods 0.000 claims description 19
- 239000000284 extract Substances 0.000 claims description 18
- 230000009466 transformation Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000012937 correction Methods 0.000 claims description 2
- 230000003287 optical effect Effects 0.000 claims description 2
- 230000004807 localization Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
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)
- Length Measuring Devices By Optical Means (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
A positioning navigation system based on binocular vision and laser fusion.A binocular information acquisition, feature extraction and matching module comprises a binocular ORB feature extraction sub-module and a binocular feature matching sub-module; the 2D laser point extraction, matching and pose estimation module comprises a 2D laser data extraction sub-module, a laser point and 2D map matching sub-module and a pose optimization sub-module; the visual SLAM pose and 2D laser pose fusion module comprises a visual SLAM pose extraction sub-module, a 2D laser pose extraction sub-module, a pose judgment sub-module, a pose output sub-module and a mapping sub-module. The invention provides a positioning navigation system based on binocular vision and laser fusion, which has the advantages of good robustness, high accuracy and strong adaptability.
Description
Technical Field
The invention belongs to the technical field of robots, and particularly relates to a robot positioning and map building system.
Background
The simultaneous localization and mapping (SLAM) technique is an important issue in the field of robot navigation, and the SLAM problem can be described as: the robot can establish a global map for the explored environment, and can use the map to estimate the position of the robot at any time. The robot freely moves in the environment through the equipment with the sensor, positions the position of the robot through the acquired information, and simultaneously constructs a map on the basis of positioning, so that the robot can be positioned and constructed simultaneously. There are two main factors that affect the solution of the SLAM problem, namely the data characteristics of the sensor and the observation data correlation, and if the robustness and accuracy of the data correlation can be improved and the utilization rate of the sensor data can be improved, the positioning precision and the mapping precision of the robot can be improved.
At present, the most mainstream sensors in the SLAM system are a camera and a laser sensor, but at present, both a visual SLAM mainly based on the camera and a laser SLAM mainly based on the laser have some problems in precision, so how to obtain more accurate data positioning by using data fusion of a plurality of sensors is a mainstream research direction of the current SLAM.
Disclosure of Invention
In order to overcome the defects of poor robustness, low accuracy and poor adaptability of the existing robot positioning and mapping system, the invention provides the robot positioning and mapping system based on binocular vision characteristics and 2D laser sensor information, which has the advantages of good robustness, high accuracy and strong adaptability.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a positioning navigation system based on binocular vision and laser fusion comprises a binocular information acquisition, feature extraction and matching module, a 2D laser point extraction, matching and pose estimation module, a vision SLAM algorithm initialization and tracking module, a local mapping module and a vision SLAM pose and 2D laser pose fusion module;
the binocular information acquisition, feature extraction and matching module comprises a binocular ORB feature extraction submodule and a binocular feature matching submodule, wherein the binocular ORB feature extraction submodule extracts ORB features in left and right images by using an ORB feature extraction algorithm by receiving pictures shot by a binocular camera in the moving process of the robot, and stores the extracted results; the binocular feature matching submodule is used for matching the feature points extracted from the left and right pictures to obtain depth information of the feature points;
the 2D laser point extraction, matching and pose estimation module comprises a 2D laser data extraction submodule, a laser point and 2D map matching submodule and a pose optimization submodule, wherein the 2D laser data extraction submodule extracts data of each frame of laser from a laser sensor and transmits the result to the laser point and 2D map matching submodule; the laser point and 2D map matching sub-module receives data from a laser sensor, performs coordinate transformation on the data, projects the data into an existing map, then matches the data with information in the map to obtain an optimal pose which may exist in the robot, uses the pose as an initial pose of the pose optimization sub-module in the next step, and calculates a better pose result after iteration for multiple times by adopting a least square iterative algorithm;
the visual SLAM pose and 2D laser pose fusion module comprises a visual SLAM pose extraction sub-module, a 2D laser pose extraction sub-module, a pose judgment sub-module, a pose output sub-module and a mapping sub-module, wherein the visual SLAM pose extraction sub-module is used for extracting the pose increment from the visual SLAM, the 2D laser pose extraction sub-module is used for extracting the pose increment from the laser SLAM, the two quantities are input into the pose judgment sub-module, the angle weight and the speed weight are introduced to judge which of the pose increment from the visual SLAM and the pose increment from the laser SLAM is closer to an actual value, then the pose output sub-module is used as a final position increment to be output to the mapping sub-module, and finally a map is updated.
The binocular vision information, the laser sensor information and the controller information are innovatively utilized, the problem of information error of a single sensor under certain conditions is solved, and the robustness of the SLAM algorithm is improved. On the basis, a map with real scale information is constructed.
The technology of simultaneous localization and map creation (SLAM) is a more classical problem in the robot field, and the SLAM problem can be described as that a robot starts to move in an unknown environment from a certain unknown position, estimates the self pose in the moving process and establishes an environment map, so that the autonomous localization and map creation of the robot are realized. Two factors affecting the SLAM system include the correlation of observed data and environmental noise, and the correctness of the ambient observation depends on good data correlation, thereby affecting the construction of an environmental map.
The technical conception of the invention is as follows: the invention provides a robot SLAM system which is a solution based on binocular vision and a 2D laser information fusion SLAM system. Firstly, estimating the pose change of the robot by using pure visual pose information; in the visual pose, extracting abundant descriptors from the image by using a high-efficiency feature extraction algorithm, and then matching by using the extracted feature points to obtain depth information of the image; then, the current position of the robot is estimated more accurately by using the visual geometry knowledge; estimating the pose change of the robot by using the pure laser pose information; in the laser SLAM, the latest laser data is matched with the existing 2D map to obtain pose estimation, and the pose is optimized by least square iteration. And then, angle weight is introduced by using pose increment respectively obtained from the laser SLAM and the visual SLAM, aiming at some problems existing in the laser SLAM and the visual SLAM in the processes of linear motion and rotary motion, so that the confidence coefficients of the visual SLAM and the laser SLAM under different conditions are improved, and the robustness of the system is improved.
The invention has the following beneficial effects: by adopting the method of fusing visual information and 2D laser information, the problem in the visual SLAM can be well solved, and the robustness of the SLAM system is improved; high accuracy and adaptability.
Drawings
Fig. 1 is a schematic representation of the system architecture of the present invention.
Fig. 2 is a system flow diagram of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 and 2, a positioning navigation system based on binocular vision and laser fusion comprises a binocular information acquisition and feature extraction and matching module 1, a 2D laser point extraction, matching and pose estimation module 2, and a vision SLAM pose and 2D laser pose fusion module 3; the binocular information acquisition, feature extraction and matching module 1 comprises: a binocular ORB feature extraction sub-module 1.1 and a binocular feature matching sub-module 1.2; the 2D laser point extraction, matching and pose estimation module 2 includes: a 2D laser data extraction sub-module 2.1, a laser point and 2D map matching sub-module 2.2 and a pose optimization sub-module 2.3; the visual SLAM pose and 2D laser pose fusion module 3 comprises: a vision SLAM pose extraction sub-module 3.1, a 2D laser pose extraction sub-module 3.2, a pose judgment sub-module 3.3, a pose output sub-module 3.4 and a mapping sub-module 3.5.
In the binocular information acquisition, feature extraction and matching module 1, a binocular ORB feature extraction submodule 1.1 extracts ORB features in left and right images by receiving pictures shot by a binocular camera in the moving process of the robot by using an ORB feature extraction algorithm, and stores the extracted results; and the binocular feature matching submodule 1.2 is used for matching the feature points extracted from the left and right pictures to obtain the depth information of the feature points.
In the 2D laser point extraction, matching and pose estimation module 2, a 2D laser data extraction submodule 2.1 extracts data of each frame of laser from a laser sensor, and transmits the result to a laser point and 2D map matching submodule 2.2; the laser point and 2D map matching sub-module receives data from the laser sensor, carries out coordinate transformation on the data and projects the data into the existing map, then matches the data with information in the map to obtain the best possible pose of the robot, the pose is used as the initial pose of the pose optimization sub-module 2.3 in the next step, and a more optimal pose result is calculated after a least square iterative algorithm is adopted for multiple iterations.
In the visual SLAM pose and 2D laser pose fusion module 3, a visual SLAM pose extraction sub-module 3.1 extracts a pose increment from a visual SLAM, a 2D laser pose extraction sub-module 3.2 extracts a pose increment from a laser SLAM, the two quantities are input into a pose judgment sub-module 3.3, angle weight and speed weight are introduced to judge which of the pose increment from the visual SLAM and the pose increment from the laser SLAM is closer to an actual value, then a pose output sub-module 3.4 outputs the pose increment as a final position increment to a mapping module pose output sub-module 3.5 by a pose output sub-module 3.4, and finally a map is updated.
The execution of the modules is described in further detail below.
In the binocular information acquisition, feature extraction and matching module 1:
the binocular ORB feature extraction submodule 1.1 is used for extracting ORB features of images after the left and right images are collected by the robot each time, and the reason that the ORB features are selected as the image features is that the ORB features have good visual angle illumination invariance, and the ORB features are used as binary visual features, have the characteristics of high extraction speed and high matching speed, and are very suitable for SLAM systems requiring real-time performance. After the post-ORB features are extracted, the extracted results are stored.
The binocular feature matching submodule 1.2 is used for matching only in the same line when searching for matching points because left and right images are subjected to polar line correction, and after the robot collects images and extracts feature points each time, absolute scales of the feature points can be calculated by using a binocular vision trigonometry, and a calculation formula is as follows:
in the above formula, f is the focal length of the camera, and B is the distance between the optical centers of the two cameras; d is parallax, namely the distance difference of the same characteristic point of the left and right pictures, and z is the depth of the characteristic point;
matching the feature points with the feature points in the previous frame or the key frame according to the information of the feature points, establishing an error function by applying a reprojection error, and finally solving the variation of the camera pose by solving the minimized error function.
In the 2D laser point extraction, matching and pose estimation module 2, a 2D laser data extraction submodule 2.1 extracts data of each frame of laser from a laser sensor and transmits the result to a laser point and 2D map matching submodule 2.2; the laser point and 2D map matching submodule receives data from a laser sensor, performs coordinate transformation on the data and then projects the data to the existing map, and firstly converts the data from the laser sensor into coordinate points under a rectangular coordinate system from a polar coordinate system:
where r is the distance information returned for each laser data and θ is the angle information for each laser beam,(s) i，x ，S i，y ) T Coordinates of each laser point which takes the laser as the center after conversion under a rectangular coordinate system;
then, the coordinate system centered on the laser is converted into the coordinate system under the world coordinate system:
wherein
Is the coordinate of the robot in the current world coordinate system, S i (delta) is the laser spot S i ＝(s i ，x，S i，y ) T The world coordinates of (a);
then matching the pose with information in a map to obtain the possible optimal pose of the robot, taking the pose as the initial pose of a pose optimization submodule in the next step, and calculating a better pose result after iteration is performed for multiple times by adopting a least square iteration algorithm;
here, an occupied grid map is used, in which each grid point is represented by a probability for each laser spot S i (δ) its probability value is obtained by a bilinear difference value with the probability values of the surrounding four integer points:
meanwhile, the partial derivative of this point is expressed as:
from the trellis probability values of each point, we can obtain a corresponding error function:
the formula is in a least square form, and finally, an optimal solution of the formula is obtained by solving the least square mode.
In the visual SLAM pose and 2D laser pose fusion module 3, the visual SLAM pose extraction sub-module 3.1 extracts pose increments from the visual SLAM, and the pose increments are recorded as
The pose extraction from 2D laser pose extraction sub-module 3.2 will then extract the pose increment from the laser SLAM, here denoted as
Consider (Δ z) in the above pose, with we only moving horizontally ca ，Δα ca ，Δβ ca ) Substantially unchanged, so the same considerations apply
Three quantities of pose changes are input into a pose judgment sub-module 3.3, and by introducing angle weights, the pose increment from the vision SLAM and the pose increment from the laser SLAM are judged to be closer to an actual value;
how to obtain the final pose estimation according to the variation of the two poses is a simple way of averaging:
the precision problem of laser and vision under some scenes is considered: the pure laser sensor only has distance information, so that self positioning failure is easily caused in a scene that the distance information is not obviously changed or basically does not change in the walking process of a gallery, and vision can acquire a large number of characteristic points in space, so that self pose change can be calculated by measuring the change of the relative position of each characteristic point even in the scene that the distance of surrounding obstacles is not obviously changed, and more accurate pose estimation is achieved; although the vision still keeps good feature tracking performance in a rectilinear scene such as a gallery, the problem of feature matching error or matching failure and the like is easily caused in the rotation process, so that the pose tracking error is caused, however, the distance corresponding to each point in the rotation process can be changed due to the fact that the laser only uses distance information, and even if only a few feature points exist, the good tracking state can still be kept in the rotation process.
Based on the advantages and disadvantages of the laser SLAM and the visual SLAM in the straight line walking and turning processes, an angle weight judgment is introduced, in the straight line walking process, in order to avoid tracking errors in the laser SLAM, the confidence coefficient of visual pose judgment can be improved, and in the turning process, data from the laser SLAM tends to be improved.
The pose increment can be finally obtained as follows:
and then the pose increment is taken as a final position increment and is output to a mapping module pose output submodule 3.5 by a pose output submodule 3.4, and finally the map is updated.
Taking the device of the robot equipped with the binocular camera and the 2D laser sensor as an example, the implementation process implemented on the device will be described in detail.
Firstly, in the binocular information acquisition and depth information acquisition module 1, feature extraction and feature matching are carried out on binocular visual information, the obtained sensor information is transmitted to a subsequent module, and the part continuously acquires and extracts data newly obtained by a sensor because the robot moves in space. And acquiring an image by using a binocular camera, sending the image to a feature extraction algorithm, storing the positions and descriptors of the extracted image features, and performing matching calculation on the depths of the feature points and storing the feature points by using the feature points of the extracted left and right pictures. Then, the feature points are matched with the feature points in the previous image, and the estimated robot pose under the vision SLAM is obtained by minimizing the reprojection error.
And then, extracting data of the laser sensor in a 2D laser point extraction, matching and pose estimation module (2), wherein in the motion process of the robot, the laser data is sent at a certain frequency in real time and is received by a 2D laser SLAM. After receiving data from a sensor, the data is firstly converted into a rectangular coordinate system with a laser coordinate as a center, then the coordinate is converted into a coordinate under a world coordinate system through the pose relation of the robot under the world coordinate system, finally, a laser beam is aligned with a map through a Gauss-Newton method, a least square problem is constructed, and a final result, namely the pose of the laser SLAM, is obtained through a general solution form of the least square method.
And finally, in the visual SLAM pose and 2D laser pose fusion module (3), the visual SLAM pose extraction sub-module and the laser SLAM pose extraction sub-module respectively extract the estimated poses from the visual SLAM and the laser SLAM, the advantages and the disadvantages of the visual SLAM and the laser SLAM are fully considered, and the angle weight is introduced, so that the confidence coefficient of the laser or visual pose estimation of the system is respectively improved under different conditions, and the mapping accuracy of the system is improved.
The embodiments described in this specification are merely illustrative of implementations of the inventive concepts, which are intended for purposes of illustration only. The scope of the present invention should not be construed as being limited to the particular forms set forth in the examples, but rather as being defined by the claims and the equivalents thereof which can occur to those skilled in the art upon consideration of the present inventive concept.
Claims (4)
1. A positioning navigation system based on binocular vision and laser fusion is characterized by comprising a binocular information acquisition, feature extraction and matching module, a 2D laser point extraction, matching and pose estimation module and a vision SLAM pose and 2D laser pose fusion module; the binocular information acquisition, feature extraction and matching module comprises a binocular ORB feature extraction submodule and a binocular feature matching submodule, wherein the binocular ORB feature extraction submodule extracts ORB features in left and right images by receiving pictures shot by a binocular camera in the moving process of the robot by using an ORB feature extraction algorithm, and stores the extracted results; the binocular feature matching submodule is used for matching the feature points extracted from the left and right pictures to obtain depth information of the feature points;
the 2D laser point extraction, matching and pose estimation module comprises a 2D laser data extraction submodule, a laser point and 2D map matching submodule and a pose optimization submodule, wherein the 2D laser data extraction submodule extracts data of each frame of laser from a laser sensor and transmits the result to the laser point and 2D map matching submodule; the laser point and 2D map matching sub-module receives data from a laser sensor, performs coordinate transformation on the data, projects the data into an existing map, then matches the data with information in the map to obtain an optimal pose which may exist in the robot, uses the pose as an initial pose of the pose optimization sub-module in the next step, and calculates a better pose result after iteration for multiple times by adopting a least square iterative algorithm;
the visual SLAM pose and 2D laser pose fusion module comprises a visual SLAM pose extraction sub-module, a 2D laser pose extraction sub-module, a pose judgment sub-module, a pose output sub-module and a mapping sub-module, wherein the visual SLAM pose extraction sub-module is used for extracting the pose increment from the visual SLAM, the 2D laser pose extraction sub-module is used for extracting the pose increment from the laser SLAM, the two quantities are input into the pose judgment sub-module, the pose increment from the visual SLAM and the pose increment from the laser SLAM are judged to be closer to an actual value by introducing angle weight, then the pose output sub-module is used as a final position increment to be output to the mapping sub-module, and finally a map is updated;
the visual SLAM pose extraction sub-module will extract the pose increment from the visual SLAM, denoted first here as
The pose extraction from 2D laser pose extraction sub-module will then extract the pose increment from the laser SLAM, here denoted as
Consider (Δ z) in the above pose only in the case of horizontal movement ca ,Δα ca ,Δβ ca ) Substantially unchanged, so the same considerations apply
Three quantities of pose changes are input into a pose judgment sub-module, and by introducing angle weight, the pose increment from the vision SLAM and the pose increment from the laser SLAM are judged to be closer to an actual value;
how to obtain the final pose estimation according to the variation of the two poses is a simple way of averaging:
an angle weight judgment is introduced, so that the confidence of the visual pose judgment can be improved in order to avoid tracking errors in the laser SLAM during straight line walking, and the data from the laser SLAM is more prone to be improved in the turning process;
the pose increment is finally obtained as follows:
and then the pose increment is taken as a final position increment and is output to a pose output submodule of the mapping module by a pose output submodule, and finally the map is updated.
2. The binocular vision and laser fusion based positioning navigation system of claim 1, wherein the binocular ORB feature extraction sub-module is configured to extract ORB features of the images after the left and right images are acquired by the robot each time, and to store the extracted ORB features after the extracted ORB features are extracted.
3. The binocular vision and laser fusion based positioning navigation system of claim 2, wherein in the binocular feature matching sub-module, since left and right images are subjected to epipolar line correction, matching is only needed to be performed in the same row when searching for matching points, and after the robot collects the images and extracts feature points each time, the descriptor matching of the feature points is utilized, the absolute scale of the feature points can be calculated by utilizing a binocular vision trigonometry, and the calculation formula is as follows:
in the above formula, f is the focal length of the camera, and B is the distance between the optical centers of the two cameras; d is parallax, namely the distance difference of the same characteristic point of the left picture and the right picture, and z is the depth of the characteristic point;
matching the feature points with the feature points in the previous frame or the key frame according to the information of the feature points, establishing an error function by applying a reprojection error, and finally solving the variation of the camera pose by solving the minimized error function.
4. The binocular vision and laser fusion based positioning navigation system of one of claims 1 to 3, wherein the 2D laser data extraction sub-module extracts data of each frame of laser light from a laser sensor and delivers the result to a laser point and 2D map matching sub-module; the laser point and 2D map matching submodule receives data from the laser sensor, performs coordinate transformation on the data and then projects the data into the existing map, and firstly converts the data from the laser sensor into coordinate points under a rectangular coordinate system from a polar coordinate system:
where r is the distance information returned for each laser data and θ is the angle information for each laser beam,(s) i,x ,S i,y ) T For each laser spot centred on the laser after conversionCoordinates under a coordinate system;
then, the coordinate system centered on the laser is converted into the coordinate system under the world coordinate system:
wherein
Is the coordinate of the robot in the current world coordinate system, S i (delta) is the laser spot S i ＝(s i,x ,S i,y ) T The world coordinates of (a);
then matching the pose with information in a map to obtain the possible optimal pose of the robot, taking the pose as the initial pose of a pose optimization submodule in the next step, and calculating a better pose result after iteration is performed for multiple times by adopting a least square iteration algorithm;
here, an occupied grid map is used, in which each grid point is represented by a probability for each laser spot S i (δ) its probability value is obtained by a bilinear difference value using the probability values of the surrounding four integer points:
meanwhile, the partial derivative of this point is expressed as:
obtaining a corresponding error function through the grid probability value of each point:
and finally, solving to obtain an optimal solution of the formula in a least square solving mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011537391.2A CN112747749B (en) | 2020-12-23 | 2020-12-23 | Positioning navigation system based on binocular vision and laser fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011537391.2A CN112747749B (en) | 2020-12-23 | 2020-12-23 | Positioning navigation system based on binocular vision and laser fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112747749A CN112747749A (en) | 2021-05-04 |
| CN112747749B true CN112747749B (en) | 2022-12-06 |
Family
ID=75646198
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011537391.2A Active CN112747749B (en) | 2020-12-23 | 2020-12-23 | Positioning navigation system based on binocular vision and laser fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112747749B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113758415A (en) * | 2021-06-30 | 2021-12-07 | 广东食品药品职业学院 | Machine vision positioning support, system and positioning method based on deep learning |
| CN114355908A (en) * | 2021-12-22 | 2022-04-15 | 无锡江南智造科技股份有限公司 | Navigation optimization method based on feature recognition |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665540A (en) * | 2018-03-16 | 2018-10-16 | 浙江工业大学 | Robot localization based on binocular vision feature and IMU information and map structuring system |
| CN109341705A (en) * | 2018-10-16 | 2019-02-15 | 北京工业大学 | Intelligent detecting robot simultaneous localization and mapping system |
| CN111045017A (en) * | 2019-12-20 | 2020-04-21 | 成都理工大学 | Method for constructing transformer substation map of inspection robot by fusing laser and vision |
| CN111337947A (en) * | 2020-05-18 | 2020-06-26 | 深圳市智绘科技有限公司 | Instant mapping and positioning method, device, system and storage medium |
| CN111966101A (en) * | 2020-08-18 | 2020-11-20 | 国以贤智能科技(上海)股份有限公司 | Turning control method, device and system for unmanned mobile device and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3032812A1 (en) * | 2016-08-04 | 2018-02-08 | Reification Inc. | Methods for simultaneous localization and mapping (slam) and related apparatus and systems |
| CN107167148A (en) * | 2017-05-24 | 2017-09-15 | 安科机器人有限公司 | Synchronous superposition method and apparatus |
| CN107796397B (en) * | 2017-09-14 | 2020-05-15 | 杭州迦智科技有限公司 | Robot binocular vision positioning method and device and storage medium |
| US11243081B2 (en) * | 2019-03-29 | 2022-02-08 | Trimble Inc. | Slam assisted INS |
| CN111595333B (en) * | 2020-04-26 | 2023-07-28 | 武汉理工大学 | Modularized unmanned vehicle positioning method and system based on visual inertia laser data fusion |
-
2020
- 2020-12-23 CN CN202011537391.2A patent/CN112747749B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665540A (en) * | 2018-03-16 | 2018-10-16 | 浙江工业大学 | Robot localization based on binocular vision feature and IMU information and map structuring system |
| CN109341705A (en) * | 2018-10-16 | 2019-02-15 | 北京工业大学 | Intelligent detecting robot simultaneous localization and mapping system |
| CN111045017A (en) * | 2019-12-20 | 2020-04-21 | 成都理工大学 | Method for constructing transformer substation map of inspection robot by fusing laser and vision |
| CN111337947A (en) * | 2020-05-18 | 2020-06-26 | 深圳市智绘科技有限公司 | Instant mapping and positioning method, device, system and storage medium |
| CN111966101A (en) * | 2020-08-18 | 2020-11-20 | 国以贤智能科技(上海)股份有限公司 | Turning control method, device and system for unmanned mobile device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112747749A (en) | 2021-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110261870B (en) | Synchronous positioning and mapping method for vision-inertia-laser fusion | |
| CN112985416B (en) | Robust positioning and mapping method and system based on laser and visual information fusion | |
| CN111795686B (en) | Mobile robot positioning and mapping method | |
| CN104732518B (en) | A kind of PTAM improved methods based on intelligent robot terrain surface specifications | |
| CN110033489B (en) | Method, device and equipment for evaluating vehicle positioning accuracy | |
| CN109191504A (en) | A kind of unmanned plane target tracking | |
| CN112785702A (en) | SLAM method based on tight coupling of 2D laser radar and binocular camera | |
| US5422828A (en) | Method and system for image-sequence-based target tracking and range estimation | |
| CN105783913A (en) | SLAM device integrating multiple vehicle-mounted sensors and control method of device | |
| CN108519102B (en) | Binocular vision mileage calculation method based on secondary projection | |
| CN112747749B (en) | Positioning navigation system based on binocular vision and laser fusion | |
| CN111665512B (en) | Ranging and mapping based on fusion of 3D lidar and inertial measurement unit | |
| US11195297B2 (en) | Method and system for visual localization based on dual dome cameras | |
| CN116222543B (en) | Multi-sensor fusion map construction method and system for robot environment perception | |
| CN114494629A (en) | Three-dimensional map construction method, device, equipment and storage medium | |
| CN115272596A (en) | Multi-sensor fusion SLAM method oriented to monotonous texture-free large scene | |
| Lin et al. | A sparse visual odometry technique based on pose adjustment with keyframe matching | |
| CN110751123A (en) | Monocular vision inertial odometer system and method | |
| CN116128966A (en) | Semantic positioning method based on environmental object | |
| CN115031718A (en) | Unmanned ship synchronous positioning and mapping method (SLAM) and system with multi-sensor fusion | |
| CN116380079A (en) | Underwater SLAM method for fusing front-view sonar and ORB-SLAM3 | |
| CN116007609A (en) | Positioning method and computing system for fusion of multispectral image and inertial navigation | |
| CN118168545A (en) | Positioning navigation system and method for weeding robot based on multi-source sensor fusion | |
| CN112762929B (en) | Intelligent navigation method, device and equipment | |
| CN117115271A (en) | Binocular camera external parameter self-calibration method and system in unmanned aerial vehicle flight process |
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 | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A positioning and navigation system based on binocular vision and laser fusion Effective date of registration: 20231213 Granted publication date: 20221206 Pledgee: Baochu sub branch of Bank of Hangzhou Co.,Ltd. Pledgor: ZHEJIANG TONGZHU TECHNOLOGY Co.,Ltd. Registration number: Y2023330003008 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right |