CN114577215A - Method, device and medium for updating feature map of mobile robot - Google Patents
Method, device and medium for updating feature map of mobile robot Download PDFInfo
- Publication number
- CN114577215A CN114577215A CN202210236656.8A CN202210236656A CN114577215A CN 114577215 A CN114577215 A CN 114577215A CN 202210236656 A CN202210236656 A CN 202210236656A CN 114577215 A CN114577215 A CN 114577215A
- Authority
- CN
- China
- Prior art keywords
- feature
- feature points
- points
- map
- point
- 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
- 238000000034 method Methods 0.000 title claims abstract description 59
- 239000000126 substance Substances 0.000 claims description 3
- 230000008569 process Effects 0.000 description 18
- 238000005516 engineering process Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000033001 locomotion Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000007704 transition Effects 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
- G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
- G01C21/3804—Creation or updating of map data
- G01C21/3807—Creation or updating of map data characterised by the type of data
- G01C21/3811—Point data, e.g. Point of Interest [POI]
-
- 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)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The embodiment of the specification discloses a method, equipment and a medium for updating a feature map of a mobile robot, wherein the method comprises the following steps: acquiring point cloud data of the mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data; respectively matching a plurality of feature points in the current feature point set with feature points in a preset initialized feature map in nearest neighbor mode; according to the matching result of nearest neighbor matching, dividing a plurality of feature points in a current feature point set into first-class feature points and second-class feature points, and according to the position information of the first-class feature points, updating the positions of corresponding feature points matched with the first-class feature points in a preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
Description
Technical Field
The present disclosure relates to the field of robotics, and in particular, to a method, an apparatus, and a medium for updating a feature map of a mobile robot.
Background
With the development of scientific technology, the application field of the mobile robot is gradually expanded, the facing working environment is more and more complex, the realization of autonomous navigation of the robot is an important step for solving the problems generated in various complex working environments, the mobile robot map construction is a key technology in the mobile robot technology, and the construction of the feature map directly concerns the accuracy and stability of the mobile robot in the subsequent positioning process.
In an application scene of an indoor mobile robot, surrounding environment features such as arcs, line segments, angular points and the like are often scanned by a single line laser radar for recording, and are stored in a feature map. However, during the movement of the mobile robot, there may be cases where redundant features are scanned, for example, features scanned from dynamic obstacles, features blocked by other obstacles, and the like. Because the scanned features may have redundant features, the updated feature map has a poor quality problem, which affects the stability and accuracy in the subsequent positioning process.
Disclosure of Invention
One or more embodiments of the present specification provide a method, an apparatus, and a medium for updating a feature map of a mobile robot, which are used to solve the following technical problems: because the scanned features may have redundant features, the updated feature map has a poor quality problem, which affects the stability in the subsequent positioning process.
One or more embodiments of the present disclosure adopt the following technical solutions:
one or more embodiments of the present specification provide a feature map updating method of a mobile robot, the method including: acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points; respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment; according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map; according to the position information of the first type of feature points acquired in advance, updating the position information of the corresponding feature points matched with the first type of feature points in the preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
Further, before performing nearest neighbor matching on the plurality of feature points in the current feature point set with the feature points in the preset initialized feature map, the method further includes: acquiring a plurality of feature points in the point cloud data of the mobile robot at the previous moment of the current moment; and performing initialization definition on each feature point in the plurality of feature points to generate the initialization feature map.
Further, the initializing definition of each feature point in the plurality of feature points specifically includes: obtaining the feature type of each feature point in the plurality of feature points, and defining the feature type of each feature point in the plurality of feature points according to the feature type of each feature point, wherein the feature type comprises a circular arc class and an angular point class; acquiring the global position coordinates of each feature point in the plurality of feature points, and defining the position information of each feature point according to the global position coordinates of each feature point; setting the theoretical observation times and the effective observation times of each feature point in the plurality of feature points as initial values; defining each feature point of the plurality of feature points as an untrusted landmark.
Further, the performing nearest neighbor matching on the plurality of feature points in the current feature point set with the feature points in the preset initialized feature map respectively specifically includes: acquiring position information of each feature point in the initialized feature map; and performing nearest neighbor matching on the plurality of feature points in the current feature point set and the position information of each feature point in the initialized feature map one by one.
Further, the classifying the plurality of feature points in the current feature point set into a first class of feature points and a second class of feature points according to the matching result of the nearest neighbor matching specifically includes: if a designated feature point exists in a plurality of feature points in the current feature point set, and the distance between the designated feature point and a corresponding feature point in the initialized feature map is smaller than a preset threshold value, judging that the matching result of the nearest neighbor matching is successful, and taking the designated feature point as a first class feature point; if a current feature point exists in a plurality of feature points in the current feature point set, and the distance between the current feature point and any one feature point in the initialized feature map is greater than or equal to a preset threshold value, determining that the matching result of the nearest neighbor matching is unsuccessful, and taking the current feature point as a second-class feature point.
Further, the updating, according to the position information of the first-class feature points acquired in advance, the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate the first feature map specifically includes: acquiring position information of each feature point in the first type of feature points as first position information; acquiring position information of each current feature point matched with each feature point in the first class of feature points in the preset initialization feature map as second position information; updating the position information of each current feature point according to the first position information and the second position information, and generating the position information of each current feature point so as to update the position information of each current feature point; and setting the effective observation times of each current feature point matched with each feature point in the first class of feature points in the preset initialized feature map as an initial value plus one so as to update the effective observation times and generate the updated effective observation times of each current feature point.
Further, after generating the updated feature map, the method further includes: generating an effective observation rate of each feature point in the updated feature map according to the ratio of the effective observation times to the theoretical observation times of each feature point in the updated feature map; judging whether feature points to be eliminated exist in the feature points in the updated feature map or not according to the effective observation times of the feature points in the updated feature map and the effective observation rate of the feature points in the updated feature map, wherein the updated effective observation times of the feature points to be eliminated are smaller than a first preset threshold value, and the effective observation rate is smaller than a second preset threshold value; and when the feature points to be removed exist in the feature points in the updated feature map, removing the feature points to be removed from the updated feature map.
Further, after generating the updated feature map, the method further includes: if a first feature point exists in each feature point in the updated feature map and the updated effective observation times are larger than or equal to a first preset threshold, updating the first feature point from a non-trusted landmark to a trusted landmark.
One or more embodiments of the present specification provide a feature map updating apparatus of a mobile robot, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points; respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment; according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map; according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance to generate an updated feature map.
One or more embodiments of the present specification provide a non-transitory computer storage medium storing computer-executable instructions configured to: acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points; respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment; according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map; according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
The embodiment of the specification adopts at least one technical scheme which can achieve the following beneficial effects: through the technical scheme, the inherent attributes are defined for each feature to initialize the feature map, the feature points in the current feature point set are divided into the matchable features and the new features through matching of the current feature point set and the initialized feature map, the Kalman filter is independently established for each feature point, updating of each feature is tracked, and the feature points are evaluated and rejected according to the historical observation effect of the features, so that the calculation memory is saved, the stability of the feature map can be greatly improved, the mapping quality of the robot is ensured, and the positioning and obstacle avoidance capabilities are improved.
Drawings
In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present specification, and for those skilled in the art, other drawings can be obtained according to the drawings without any creative effort. In the drawings:
fig. 1 is a schematic flowchart of a method for updating a feature map of a mobile robot according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of another method for updating a feature map of a mobile robot according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a feature map updating apparatus for a mobile robot according to an embodiment of the present disclosure.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the present specification, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present specification without any creative effort shall fall within the protection scope of the present specification.
With the development of scientific technology, the application field of the mobile robot is gradually expanded, the facing working environment is more and more complex, the realization of autonomous navigation of the robot is an important step for solving the problems generated in various complex working environments, the mobile robot map construction is a key technology in the mobile robot technology, and the construction of the feature map directly concerns the accuracy and stability of the mobile robot in the subsequent positioning process.
In an application scene of an indoor mobile robot, surrounding environment features such as arcs, line segments, angular points and the like are often scanned by a single line laser radar for recording, and are stored in a feature map. However, during the movement of the mobile robot, there may be cases where redundant features are scanned, for example, features scanned from dynamic obstacles, features blocked by other obstacles, and the like. Because the scanned features may have redundant features, the updated feature map has a poor quality problem, which affects the stability and accuracy in the subsequent positioning process.
The embodiment of the specification provides a method for updating a feature map of a mobile robot, and an execution main body can be a server or any equipment with data processing capacity. Fig. 1 is a schematic flow chart of a method for updating a feature map of a mobile robot according to an embodiment of the present disclosure, and as shown in fig. 1, the method mainly includes the following steps:
step S101, point cloud data of the mobile robot at the current moment is obtained, and a current feature point set is generated according to a plurality of feature points in the point cloud data.
In one embodiment of the present specification, the point cloud data of the mobile robot at the current time is obtained by the laser radar, and the point cloud data obtained by the laser radar is different at different times along with the movement of the robot due to the mobile robot being in a mobile state. After point cloud data of the current moment are obtained, after features are extracted from the point cloud data, a plurality of feature points are generated, and the plurality of feature points in the point cloud data form a current feature point set.
And step S102, respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and the feature points in the preset initialized feature map.
And the preset initialization feature map is the initialization feature map corresponding to the last moment of the current moment.
Respectively matching a plurality of feature points in the current feature point set with feature points in a preset initialization feature map in a nearest neighbor manner, and the method further comprises the following steps: acquiring a plurality of feature points in point cloud data of a mobile robot at the previous moment of the current moment; initializing and defining each feature point in the plurality of feature points, and generating the initialized feature map, specifically comprising: the method comprises the steps of obtaining a feature type of each feature point in a plurality of feature points, and defining the feature type of each feature point in the plurality of feature points according to the feature type of each feature point, wherein the feature type comprises a circular arc type and an angular point type; acquiring the global position coordinates of each feature point in the plurality of feature points, and defining the position information of each feature point according to the global position coordinates of each feature point; setting the theoretical observation times and the effective observation times of each feature point in the plurality of feature points as initial values; defining each feature point of the plurality of feature points as an untrusted landmark.
In an embodiment of the present specification, the feature map needs to be initialized, and an initialized feature map is generated, where the feature map is present by default, and only the feature map needs to be updated. However, there may be a case where there is no feature map, and when there is no feature map, the feature point set of the first frame point cloud may be used for initialization. It should be noted that the initialization is to initially define the inherent attribute of each feature, and the attribute of a feature not only indicates its distinction from other features, but also includes its trend and performance in multiple historical observations. The method comprises the steps of obtaining point cloud data of the mobile robot at the previous moment of the current moment, extracting features from the point cloud data to obtain a plurality of feature points, performing initialization definition on each feature point of the plurality of feature points, and generating the initialization feature map.
In one embodiment of the present specification, the initially defining each of the plurality of feature points includes defining a feature type, defining location information, defining a number of theoretical observations, a number of valid observations, and whether the feature point is a trusted landmark. The position information may also be represented by global coordinates.
In an embodiment of the present specification, a feature type of each feature point is obtained, where the feature type includes a circular arc and a corner, and each feature is defined according to the feature type of each feature point. For example, feature point a is a circular arc class, and feature point B is a corner class. The matching and updating are conveniently carried out according to the feature points of the same type in the subsequent matching and updating process, and the operation efficiency is improved.
In one embodiment of the present specification, a global coordinate, that is, position information, of each feature point is obtained, and the global coordinate is recorded in a position information attribute of each feature point, so as to be used in subsequent updating and matching. In addition, it should be noted that, if a certain feature point in the detection range of the laser radar should be detected theoretically, the number of theoretical observation times is incremented by one, and similarly, the effective observation times refers to that when a feature is successfully matched or associated with a feature in a new feature point set, it is proved that the feature point is detected in a new laser scan, and it is stated that an effective observation is performed, and the number of effective observation times is incremented by one. In addition, in the actual observation process, the occlusion of other obstacles and the failure of feature extraction caused by the incapability of meeting the requirements of the feature points of the observation sample and other special conditions may cause the situation that the feature points which are detected are not detected, so that the effective observation degree of the feature points can be represented by setting an effective observation rate, wherein the effective observation rate is the ratio of the effective observation times to the theoretical observation times. Finally, a matchable mark of each feature point is needed to be set, and the matchable mark is used for judging whether the feature point is used as a mark of a trusted landmark. If the feature point is a tag of a trusted landmark, the matchable tag is set to 1, otherwise, to 0.
Since in one embodiment of the present specification, the initial definition is performed by the first feature point set, that is, the feature points scanned for the first time, the theoretical observation times, the effective observation times, and the effective observation rate are all set to 1. The feature points in the feature point set obtained by initial scanning cannot be judged whether to be a trusted landmark or not, and can be determined only through multiple scanning and detection, so that the matchable flag can be set to 0.
Specifically, the nearest neighbor matching is performed on a plurality of feature points in the current feature point set and feature points in a preset initialization feature map, and specifically includes: acquiring position information of each feature point in the initialized feature map; and performing nearest neighbor matching on a plurality of feature points in the current feature point set and the position information of each feature point in the initialized feature map one by one.
In one embodiment of the present specification, position information of each feature point in the initialized feature map is acquired, and a plurality of feature points in the current feature point set are acquired. And performing nearest neighbor matching on a plurality of feature points in the current feature point set and each feature point in the initialization map one by one. It should be noted that nearest neighbor matching refers to determining whether feature points meeting the distance requirement exist in two feature point sets by a nearest neighbor algorithm. For example, there are 10 feature points in the initialized feature map, there are 5 feature points in the current feature point set, the 5 feature points in the current feature point set are taken out first, and are respectively subjected to nearest neighbor matching with the 10 feature points in the initialized feature map, whether there is a feature point closest to the first feature point in the 10 feature points in the initialized feature map is determined, and by analogy, nearest neighbor matching is performed on the 5 feature points in the current feature point set.
And step S103, according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into a first class of feature points and a second class of feature points.
The feature points in the first class of feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class of feature points are not matched with the feature points in the initialized feature map;
specifically, according to the matching result of nearest neighbor matching, classifying a plurality of feature points in the current feature point set into a first class of feature points and a second class of feature points, specifically including: if the designated feature points exist in a plurality of feature points in the current feature point set, and the distance between the designated feature points and the corresponding feature points in the initialized feature map is smaller than a preset threshold value, judging that the matching result of nearest neighbor matching is successful, and taking the designated feature points as first class feature points; if the current feature point exists in a plurality of feature points in the current feature point set, and the distance between the current feature point and any one feature point in the initialized feature map is larger than or equal to a preset threshold value, judging that the matching result of nearest neighbor matching is unsuccessful, and taking the current feature point as a second class feature point.
In an embodiment of the present specification, if after performing nearest neighbor matching on a plurality of feature points in a current feature point set and feature points in a preset initialization feature map, a specified feature point exists in the plurality of feature points in the current feature point set, and a distance between the specified feature point and a corresponding feature point in the initialization feature map is smaller than a preset threshold, it is determined that a matching result of the nearest neighbor matching is successful, so that the specified feature point is taken as a first-class feature point. For example, after nearest neighbor matching, there are feature point a1, feature point B1, and feature point C1 in the initialized feature map, which are respectively matched with feature point a, feature point B, and feature point C, that is, the distance between feature point a and feature point a1 is less than the preset distance, the distance between feature point B and feature point B1 is less than the preset distance, and the distance between feature point C and feature point C1 is less than the preset distance, where the preset distance may be set to 0.1 meter, and thus, feature point a, feature point B, and feature point C are feature points of the first type.
If the current feature point exists in a plurality of feature points in the current feature point set, and the distance between the current feature point and any one feature point in the initialized feature map is larger than or equal to a preset threshold value, judging that the matching result of nearest neighbor matching is unsuccessful, and taking the current feature point as a second class feature point. Continuing with the above example, no feature point matching with feature point D and feature point E is found in the initialized feature map, that is, the distance between feature point D and each feature point in the initialized feature map is greater than or equal to the preset threshold, and similarly, the distance between feature point E and each feature point in the initialized feature map is greater than or equal to the preset threshold, so that feature point D and feature point E are taken as the second class of feature points, and may also be defined as new feature points.
And step S104, updating the position information of the corresponding characteristic points matched with the first-class characteristic points in a preset initialized characteristic map according to the position information of the first-class characteristic points acquired in advance, and generating the first characteristic map.
Specifically, according to the position information of the first-class feature points acquired in advance, in a preset initialized feature map, the position information of the corresponding feature points matched with the first-class feature points is updated, and a first feature map is generated, which specifically includes: acquiring position information of each feature point in the first type of feature points as first position information; acquiring position information of each current feature point matched with each feature point in the first class of feature points in a preset initialized feature map as second position information; updating the position information of each current characteristic point according to the first position information and the second position information to generate the position information of each current characteristic point so as to update the position information of each current characteristic point; and setting the effective observation times of each current characteristic point matched with each characteristic point in the first class of characteristic points in the preset initialized characteristic map as an initial value plus one so as to update the effective observation times and generate the updated effective observation times of each current characteristic point.
In an embodiment of the present specification, after determining the matched feature point and the new feature point in the current feature point set, the attribute information of the matched feature point in the initialized feature map needs to be updated. First, the location information is updated. Acquiring position information of each feature point in the first type of feature points as first position information; and acquiring the position information of each current feature point matched with each feature point in the first class of feature points in a preset initialization feature map as second position information. Because the laser radar has scanning errors in the scanning process, in order to increase the accuracy of the position information of the feature points, the position information of each current feature point is updated through the first position information and the second position information corresponding to adjacent moments, and the position information of each current feature point is generated so as to update the position information of each current feature point. In addition, the existence of the feature point matched with the feature point in the initialized feature map in the current feature point set indicates that the feature point is actually scanned twice, so that the effective observation times of each current feature point matched with each feature point in the first class of feature points in the initialized feature map which is set in advance are set as the initial values plus one, so that the effective observation times are updated, and the updated effective observation times of each current feature point are generated.
It should be noted that, the position information here may be represented by using a state quantity, and in the process of initializing the feature map and generating the initialized feature map, a kalman filter is established for each feature in the initialized feature map, that is, the kalman filter is established separately for the state quantity of each feature to track, so as to implement feature update for each feature point.
And step S105, adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
In an actual application process, the feature points scanned by the mobile robot at the previous time and the current time are different, and there may be newly scanned feature points. It is therefore necessary to add the newly scanned feature points to the first feature map.
In an embodiment of the present specification, position information of each feature point in the second class of feature points is obtained, each feature point in the second class of feature points is added to the first feature map according to the position information of each feature point in the second class of feature points, and attribute definition is performed on all feature points in the second class of feature points according to an initialization process for initializing the feature map.
After the updated feature map is generated, the method further comprises the following steps: generating an effective observation rate of each feature point in the updated feature map according to the ratio of the effective observation times to the theoretical observation times of each feature point in the updated feature map; judging whether feature points to be eliminated exist in the feature points in the updated feature map or not according to the effective observation times of the feature points in the updated feature map and the effective observation rates of the feature points in the updated feature map, wherein the effective observation times of the updated feature points to be eliminated are smaller than a first preset threshold value, and the effective observation rates are smaller than a second preset threshold value; and when the feature points to be removed exist in the feature points in the updated feature map, removing the feature points to be removed from the updated feature map.
In the process of drawing a mobile robot, the observation effect of some features may be very poor, and the reasons include: features detected from dynamic obstacles or blocked by other obstacles, and other special situations such as sudden increase of laser radar error, error in feature identification, and disorder of robot positioning may generate redundant features in the feature map. To control the infinite expansion of the feature map, it is necessary to eliminate the features that are erroneously recognized, while avoiding the elimination of features that have already become stable but perform poorly in a short time.
In an embodiment of the present specification, feature points with identification errors need to be removed from the updated feature map, a ratio between effective observation times and theoretical observation times of each feature point in the updated feature map is calculated, and an effective observation rate of each feature point in the updated feature map is generated. And when the effective observation times of the feature points in the replaced feature map are smaller than a first preset threshold and the effective observation rate is smaller than a second preset threshold, taking the feature points as the feature points to be removed, and removing the feature points from the updated feature map. It should be noted that the first preset threshold may be set to 20, and the second preset threshold may be set to 0.5.
After the updated feature map is generated, the method further comprises the following steps: if a first feature point exists in each feature point in the updated feature map and the updated effective observation times are larger than or equal to a first preset threshold, updating the first feature point from a non-trusted landmark to a trusted landmark.
After the characteristic points with the wrong identification are removed, if a first characteristic point exists in the updated characteristic map and the effective observation frequency of the first characteristic point is greater than or equal to a first preset threshold value, the first characteristic point is updated from a non-trusted landmark to a trusted landmark, and the matchable mark is set to be 1. That is, the observation of the feature is deemed to be reliable enough to be a trusted landmark for use by subsequent positioning algorithms.
Through the technical scheme, the inherent attributes are defined for each feature to initialize the feature map, the feature points in the current feature point set are divided into the matchable features and the new features through matching of the current feature point set and the initialized feature map, the Kalman filter is independently established for the features, updating of each feature is tracked, and the feature points are evaluated and removed according to the historical observation effect of the features, so that the calculation memory is saved, the stability of the feature map is greatly improved, the image construction quality of the robot is ensured, and the positioning and obstacle avoidance capabilities are improved.
Fig. 2 is a schematic flow chart of the method for updating a feature map of another mobile robot provided in the embodiment of the present specification, and the method is mainly divided into three parts, namely initialization of the feature map, updating of the feature map, and evaluation and removal of features. As shown in fig. 2, first, it is determined whether the feature map has been initially defined, and if the feature map has not been initially defined, the feature map is initialized. If the feature map is initially defined, traversing each feature point in the current feature point set, and performing the following operations: and taking out one feature point in the current feature point set, and carrying out nearest neighbor matching on the feature point in the initialized feature map. And judging whether the feature point is a new feature point according to the matching result, if the feature point matched with the feature point exists in the initialized feature map, indicating that the feature point is not the new feature point, performing Kalman update on the feature of the feature point, judging whether the feature requirement of a trusted landmark is met, and if the feature requirement of the trusted landmark is met, marking the feature point as the trusted landmark. And if the feature points matched with the initialized feature map do not exist in the initialized feature map, expanding the feature points into new features. And traversing all the feature points in the current feature point set according to the method, and after the traversal is finished, removing the low-quality features to generate an updated feature map.
The method comprises the following steps of initializing a feature map, updating the feature map, evaluating and eliminating features, wherein the specific processes of each part are as follows: firstly, initializing a feature map, defining SkAs a characteristic map of time k, CkExtracting feature for current point cloud to obtain feature point set, and when feature map does not exist, using first frame point cloud feature point set to make SkInitialization is performed. The initialization is to initially define the inherent attribute of each feature, the attribute of a certain feature not only indicates the difference from other features, but also includes the change trend and performance of the feature in multiple historical observations, in this embodiment, 6 attributes are defined for a single feature, respectively as follows:
(1) type (b). The characteristic types are divided into circular arcs and angular points, and the two characteristics are distinguished and processed during updating and matching, so that the operation efficiency can be greatly improved.
(2) Global coordinates. The global position of a current feature point of a certain feature is recorded, the current state quantity of the feature is the same, the coordinate is used for subsequent updating and serving as a map matching feature, the feature parameterized expression is embodied, and the feature point coordinate is updated all the time along with the accumulation of multiple observations of the certain feature.
(3) Number of theoretical observations NA. If a feature is within the detection range of the lidar, it should theoretically be successfully detected, NAThe value is incremented by one.
(4) Number of effective observations NV. And (3) every time a certain feature is successfully associated with the feature points in the new feature point set, the feature points prove to be successfully observed in the laser scanning, and N is the numberVThe value is incremented by one.
(5) Effective rate of observation rV. In the actual observation process, the failure of feature extraction caused by the shielding of other obstacles and the incapability of observing sample points and other special conditions can cause the failure of observing certain theoretical observable features, and r isVThe method can be used for representing the effective observation degree of a certain characteristic, and the calculation formula is as follows:
(6) the tag T can be matched. It is determined whether the feature can be used as a marker for a trusted landmark, which has a value of "0" or "1".
All feature points in the first feature point set are respectively used as independent features to record the type and global coordinates of the feature points, and N is the same as the NA、NVAnd rVSetting the data to be 1 and 0, and finishing the initialization of the feature map.
The second is the updating of the feature map. Defining a characteristic global coordinate as a state quantity
The upper subscript and the lower subscript respectively represent a characteristic serial number and time, and the Kalman filter motion and observation model is set as follows:
in the formula (I), the compound is shown in the specification,
for a new observed quantity of a certain feature, only static features are observed in the embodiment, so that the features are considered to be static, the expressions of the observed quantity and the state quantity are consistent, and the state transition matrix FBAnd observation matrix HBAre all identity matrices, wBAnd vBRespectively representing process noise and observation noise which are zero mean Gaussian white noise and are independent from each other, and the characteristic observation quality is influenced by laser radar error, characteristic fitting error and positioning error, so QBAnd RBCan be adjusted by a plurality of experiments.
The state quantity can be updated only when a certain characteristic is newly observed according to the analysis of the filter model. One-by-one extracting current feature point set feature point, making nearest neighbor matching with feature state quantity in feature map, i.e. feature point and state quantity with nearest distance in feature map form point pair, defining two points distance in said point pair as dmatchThe conditions for successful matching of the two points are determined as follows:
dmatch＜0.1m
after the point pair is identified to be successfully matched, the point pair is regarded as an effective observation, and the characteristic NVAdding one to the value, taking the characteristic point as a new observed quantity to update the corresponding state quantity, wherein the Kalman filter updating formula of a single characteristic is as follows:
in the formula (I), the compound is shown in the specification,
and
the covariance matrices are estimated a priori and the state quantities a priori respectively,
to know
The a posteriori estimated covariance matrix and kalman gain are represented, respectively. Through the Kalman filtering process, along with the gradual accumulation of observation on a certain characteristic, the covariance matrix is also converged continuously, and the state quantity tends to be stable.
Using N in the feature map update processA、NVAnd rVThree parameters evaluate the historical observation performance of a certain characteristic, and the observation effect of some characteristics may be very poor in the mobile robot mapping process, for reasons including:
(1) the characteristic is detected from a dynamic obstacle;
(2) this feature is obscured by other obstacles;
(3) other special situations, such as sudden increase of laser radar error, wrong feature identification, and disorder of robot positioning, may generate redundant features in the feature map, and all three parameter values of such features are very low.
In order to control the infinite expansion of the feature map, the features with wrong identification need to be eliminated, and meanwhile, the features which tend to be stable and have poor performance in a short time need to be avoided being eliminated, and the following conditions are set by combining the analysis:
if a certain feature simultaneously meets the two conditions, the feature is removed from the feature map. When a certain feature satisfies the following condition:
NV≥20
i.e., it is assumed that the observation of the feature is reliable enough to serve as a trusted landmark for use in subsequent localization algorithms, which may match the tag T to a "1".
An embodiment of the present specification further provides a feature map updating apparatus for a mobile robot, and fig. 3 is a schematic structural diagram of the feature map updating apparatus for a mobile robot provided in the embodiment of the present specification, and as shown in fig. 3, the apparatus includes:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to: acquiring point cloud data of the mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points; respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment; according to the matching result of nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map; according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in a preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
Embodiments of the present description also provide a non-volatile computer storage medium storing computer-executable instructions configured to: acquiring point cloud data of the mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points; respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment; according to the matching result of nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map; according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in a preset initialized feature map to generate a first feature map; and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance to generate an updated feature map.
All the embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the apparatus, device, and non-volatile computer storage medium embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to the partial description of the method embodiments for relevant points.
The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The above description is merely one or more embodiments of the present disclosure and is not intended to limit the present disclosure. Various modifications and alterations to one or more embodiments of the present description will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement or the like made within the spirit and principle of one or more embodiments of the present specification should be included in the scope of the claims of the present specification.
Claims (10)
1. A method for updating a feature map of a mobile robot, the method comprising:
acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points;
respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment;
according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map;
according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate a first feature map;
and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
2. The method according to claim 1, wherein before performing nearest neighbor matching between the plurality of feature points in the current feature point set and the feature points in the preset initialization feature map, the method further comprises:
acquiring a plurality of feature points in the point cloud data of the mobile robot at the previous moment of the current moment;
and performing initialization definition on each feature point in the plurality of feature points to generate the initialization feature map.
3. The method according to claim 2, wherein the initially defining each of the plurality of feature points specifically includes:
obtaining the feature type of each feature point in the plurality of feature points, and defining the feature type of each feature point in the plurality of feature points according to the feature type of each feature point, wherein the feature type comprises a circular arc class and an angular point class;
acquiring the global position coordinates of each feature point in the plurality of feature points, and defining the position information of each feature point according to the global position coordinates of each feature point;
setting the theoretical observation times and the effective observation times of each feature point in the plurality of feature points as initial values;
defining each feature point of the plurality of feature points as an untrusted landmark.
4. The method according to claim 3, wherein the performing nearest neighbor matching between the plurality of feature points in the current feature point set and the feature points in a preset initialization feature map specifically comprises:
acquiring position information of each feature point in the initialized feature map;
and performing nearest neighbor matching on the plurality of feature points in the current feature point set and the position information of each feature point in the initialized feature map one by one.
5. The method according to claim 1, wherein the classifying the plurality of feature points in the current feature point set into a first class of feature points and a second class of feature points according to the matching result of the nearest neighbor matching specifically comprises:
if a designated feature point exists in a plurality of feature points in the current feature point set, and the distance between the designated feature point and a corresponding feature point in the initialized feature map is smaller than a preset threshold value, judging that the matching result of the nearest neighbor matching is successful, and taking the designated feature point as a first class feature point;
if a current feature point exists in a plurality of feature points in the current feature point set, and the distance between the current feature point and any one feature point in the initialized feature map is greater than or equal to a preset threshold value, determining that the matching result of the nearest neighbor matching is unsuccessful, and taking the current feature point as a second-class feature point.
6. The method according to claim 3, wherein the updating, according to the position information of the first type of feature points acquired in advance, the position information of the corresponding feature points matched with the first type of feature points in the preset initialized feature map to generate the first feature map specifically comprises:
acquiring position information of each feature point in the first type of feature points as first position information;
acquiring position information of each current feature point matched with each feature point in the first class of feature points in the preset initialization feature map as second position information;
updating the position information of each current feature point according to the first position information and the second position information, and generating the position information of each current feature point so as to update the position information of each current feature point;
and setting the effective observation times of each current feature point matched with each feature point in the first class of feature points in the preset initialized feature map as initial values and adding one to update the effective observation times so as to generate the updated effective observation times of each current feature point.
7. The method of claim 6, wherein after generating the updated feature map, the method further comprises:
generating an effective observation rate of each feature point in the updated feature map according to the ratio of the effective observation times to the theoretical observation times of each feature point in the updated feature map;
judging whether feature points to be eliminated exist in the feature points in the updated feature map or not according to the effective observation times of the feature points in the updated feature map and the effective observation rate of the feature points in the updated feature map, wherein the updated effective observation times of the feature points to be eliminated are smaller than a first preset threshold value, and the effective observation rate is smaller than a second preset threshold value;
and when the feature points to be removed exist in the feature points in the updated feature map, removing the feature points to be removed from the updated feature map.
8. The method of claim 6, wherein after generating the updated feature map, the method further comprises:
if a first feature point exists in each feature point in the updated feature map and the updated effective observation times are larger than or equal to a first preset threshold, updating the first feature point from a non-trusted landmark to a trusted landmark.
9. A feature map updating apparatus of a mobile robot, characterized by comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points;
respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment;
according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map;
according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate a first feature map;
and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
10. A non-transitory computer storage medium storing computer-executable instructions configured to:
acquiring point cloud data of a mobile robot at the current moment, and generating a current feature point set according to a plurality of feature points in the point cloud data, wherein the point cloud data comprises a plurality of feature points;
respectively carrying out nearest neighbor matching on a plurality of feature points in the current feature point set and feature points in a preset initialized feature map, wherein the preset initialized feature map is an initialized feature map corresponding to the last moment of the current moment;
according to the matching result of the nearest neighbor matching, dividing a plurality of feature points in the current feature point set into first class feature points and second class feature points, wherein the feature points in the first class feature points are respectively matched with the corresponding feature points in the initialized feature map, and the feature points in the second class feature points are not matched with the feature points in the initialized feature map;
according to the position information of the first-class feature points acquired in advance, updating the position information of the corresponding feature points matched with the first-class feature points in the preset initialized feature map to generate a first feature map;
and adding each feature point in the second type of feature points to the first feature map according to the position information of the second type of feature points acquired in advance, and generating an updated feature map.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210236656.8A CN114577215B (en) | 2022-03-10 | 2022-03-10 | Method, equipment and medium for updating characteristic map of mobile robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210236656.8A CN114577215B (en) | 2022-03-10 | 2022-03-10 | Method, equipment and medium for updating characteristic map of mobile robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114577215A true CN114577215A (en) | 2022-06-03 |
| CN114577215B CN114577215B (en) | 2023-10-27 |
Family
ID=81780387
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210236656.8A Active CN114577215B (en) | 2022-03-10 | 2022-03-10 | Method, equipment and medium for updating characteristic map of mobile robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114577215B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101000507A (en) * | 2006-09-29 | 2007-07-18 | 浙江大学 | Method for moving robot simultanously positioning and map structuring at unknown environment |
| CN103268729A (en) * | 2013-05-22 | 2013-08-28 | 北京工业大学 | Mobile robot cascading type map creating method based on mixed characteristics |
| CN104537709A (en) * | 2014-12-15 | 2015-04-22 | 西北工业大学 | Real-time three-dimensional reconstruction key frame determination method based on position and orientation changes |
| CN107491071A (en) * | 2017-09-04 | 2017-12-19 | 中山大学 | A kind of Intelligent multi-robot collaboration mapping system and its method |
| US20180299557A1 (en) * | 2017-04-17 | 2018-10-18 | Baidu Online Network Technology (Beijing) Co., Ltd | Method and apparatus for updating maps |
| CN108917761A (en) * | 2018-05-07 | 2018-11-30 | 西安交通大学 | A kind of accurate positioning method of unmanned vehicle in underground garage |
| CN111351493A (en) * | 2018-12-24 | 2020-06-30 | 上海欧菲智能车联科技有限公司 | Positioning method and system |
| CN111780763A (en) * | 2020-06-30 | 2020-10-16 | 杭州海康机器人技术有限公司 | Visual positioning method and device based on visual map |
| US20210223046A1 (en) * | 2018-12-13 | 2021-07-22 | Goertek Inc. | Method and device for extracting key frames in simultaneous localization and mapping and smart device |
| CN113701739A (en) * | 2020-05-22 | 2021-11-26 | 杭州海康机器人技术有限公司 | Map updating method and device |
| CN114092898A (en) * | 2020-07-31 | 2022-02-25 | 华为技术有限公司 | Target object sensing method and device |
| CN114114316A (en) * | 2021-10-19 | 2022-03-01 | 山东新一代信息产业技术研究院有限公司 | Method for efficiently and accurately identifying arc features |
-
2022
- 2022-03-10 CN CN202210236656.8A patent/CN114577215B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101000507A (en) * | 2006-09-29 | 2007-07-18 | 浙江大学 | Method for moving robot simultanously positioning and map structuring at unknown environment |
| CN103268729A (en) * | 2013-05-22 | 2013-08-28 | 北京工业大学 | Mobile robot cascading type map creating method based on mixed characteristics |
| CN104537709A (en) * | 2014-12-15 | 2015-04-22 | 西北工业大学 | Real-time three-dimensional reconstruction key frame determination method based on position and orientation changes |
| US20180299557A1 (en) * | 2017-04-17 | 2018-10-18 | Baidu Online Network Technology (Beijing) Co., Ltd | Method and apparatus for updating maps |
| CN107491071A (en) * | 2017-09-04 | 2017-12-19 | 中山大学 | A kind of Intelligent multi-robot collaboration mapping system and its method |
| CN108917761A (en) * | 2018-05-07 | 2018-11-30 | 西安交通大学 | A kind of accurate positioning method of unmanned vehicle in underground garage |
| US20210223046A1 (en) * | 2018-12-13 | 2021-07-22 | Goertek Inc. | Method and device for extracting key frames in simultaneous localization and mapping and smart device |
| CN111351493A (en) * | 2018-12-24 | 2020-06-30 | 上海欧菲智能车联科技有限公司 | Positioning method and system |
| CN113701739A (en) * | 2020-05-22 | 2021-11-26 | 杭州海康机器人技术有限公司 | Map updating method and device |
| CN111780763A (en) * | 2020-06-30 | 2020-10-16 | 杭州海康机器人技术有限公司 | Visual positioning method and device based on visual map |
| WO2022002039A1 (en) * | 2020-06-30 | 2022-01-06 | 杭州海康机器人技术有限公司 | Visual positioning method and device based on visual map |
| CN114092898A (en) * | 2020-07-31 | 2022-02-25 | 华为技术有限公司 | Target object sensing method and device |
| CN114114316A (en) * | 2021-10-19 | 2022-03-01 | 山东新一代信息产业技术研究院有限公司 | Method for efficiently and accurately identifying arc features |
Non-Patent Citations (4)
| Title |
|---|
| LIN HUICAN ET AL.: "3D simultaneous localization and mapping for mobile robot based on VSLAM", JOURNAL OF COMPUTER APPLICATIONS, vol. 37, no. 10 * |
| 宋金泽等: "一种改进的RRT路径规划算法", 电子学报, vol. 38, no. 2 * |
| 张峻宁等: "一种自适应特征地图匹配的改进VSLAM算法", 自动化学报, vol. 45, no. 3 * |
| 艾青林;余杰;胡克用;陈琦;: "基于ORB关键帧匹配算法的机器人SLAM实现", 机电工程, no. 05 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114577215B (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107123164B (en) | Three-dimensional reconstruction method and system for keeping sharp features | |
| WO2020134082A1 (en) | Path planning method and apparatus, and mobile device | |
| US8199977B2 (en) | System and method for extraction of features from a 3-D point cloud | |
| CN110858076A (en) | Equipment positioning and grid map construction method and mobile robot | |
| CN112880694B (en) | Method for determining the position of a vehicle | |
| WO2022188663A1 (en) | Target detection method and apparatus | |
| EP3690707A1 (en) | Post-processing method and device for detecting lanes to plan drive path of autonomous vehicle by using segmentation score map and clustering map | |
| JP5776332B2 (en) | Map processing method and program, and robot system | |
| Senarathne et al. | Incremental algorithms for Safe and Reachable Frontier Detection for robot exploration | |
| CN109410183B (en) | Plane extraction method, system and device based on point cloud data and storage medium | |
| CN112037268B (en) | Environment sensing method based on probability transfer model in dynamic scene | |
| CN108564604B (en) | Binocular vision stereo matching method and device based on plane constraint and triangulation | |
| CN111354022A (en) | Target tracking method and system based on kernel correlation filtering | |
| CN110736456A (en) | Two-dimensional laser real-time positioning method based on feature extraction in sparse environment | |
| CN110335282B (en) | Contour line segment feature extraction method based on grids | |
| Sun et al. | Oriented point sampling for plane detection in unorganized point clouds | |
| CN111045433A (en) | Obstacle avoidance method of robot, robot and computer readable storage medium | |
| CN114577215A (en) | Method, device and medium for updating feature map of mobile robot | |
| US20200057905A1 (en) | Point group data processing device, point group data processing method, point group data processing program, vehicle control device, and vehicle | |
| WO2023072269A1 (en) | Object tracking | |
| CN114627395B (en) | Multi-rotor unmanned aerial vehicle angle analysis method, system and terminal based on nested targets | |
| US20230350422A1 (en) | Information processing apparatus, information processing method, and information processing program | |
| CN112017290B (en) | Conical surface fitting method and device, computer equipment and storage medium | |
| CN114511803A (en) | Target occlusion detection method for visual tracking task | |
| Jelinek | Vector maps in mobile robotics |
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 |