CN114529735A - Edge starting point extraction method and related device - Google Patents
Edge starting point extraction method and related device Download PDFInfo
- Publication number
- CN114529735A CN114529735A CN202111654987.5A CN202111654987A CN114529735A CN 114529735 A CN114529735 A CN 114529735A CN 202111654987 A CN202111654987 A CN 202111654987A CN 114529735 A CN114529735 A CN 114529735A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- point
- local map
- determining
- core
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a method for extracting an edge starting point and a related device, wherein the method comprises the following steps: the sweeping robot acquires an original local map; determining point cloud attributes occupying grids in an original local map, and removing the point cloud attributes as occupying grids of outliers to obtain a preprocessed local map, wherein the point cloud attributes occupying the grids comprise core points, outliers and boundary points; and performing edge starting point extraction operation on the preprocessed local map to obtain a target edge starting point. The method provided by the application can avoid the influence of the interferent on the accuracy and efficiency of the edgewise starting point extraction algorithm.
Description
Technical Field
The application relates to the technical field of sweeping robots, in particular to a method for extracting a starting point along an edge and a related device.
Background
With the development of science and technology, the sweeping robot gradually walks into the visual field of the public, and provides a great deal of convenience for the daily life of people by virtue of the characteristics of intelligence and flexibility. Sweeping is one of the main tasks of a sweeper, which typically requires mapping of the home environment before it can perform its sweeping task. Edge-based mapping is the most commonly used mapping method at present. The selection of the edgewise starting point directly influences the efficiency and performance of edgewise drawing, so that the selection of the edgewise starting point becomes a large research hot spot in the field of sweeping robot robots.
Disclosure of Invention
The method mainly solves the technical problem of providing the method and the related device for extracting the starting point along the edge, and the method and the related device can avoid the influence of an interfering object on the accuracy and the efficiency of the algorithm for extracting the starting point along the edge.
In order to solve the technical problem, the application adopts a technical scheme that: provided is an edge starting point extraction method, including:
the sweeping robot acquires an original local map;
determining point cloud attributes of occupied grids in the original local map, and removing the point cloud attributes as occupied grids of outliers to obtain a preprocessed local map, wherein the point cloud attributes of the occupied grids comprise a core point, the outliers and a boundary point;
and performing edge starting point extraction operation on the preprocessed local map to obtain a target edge starting point.
In order to solve the technical problem, the other technical scheme adopted by the application is as follows: providing a sweeping robot comprising a processor and a memory coupled to the processor; wherein the content of the first and second substances,
the memory is used for storing a computer program;
the processor is configured to run the computer program to perform the method as described in any of the above.
In order to solve the above technical problem, the present application adopts another technical solution: there is provided a computer readable storage medium storing a computer program executable by a processor for implementing a method as claimed in any one of the above.
The beneficial effect of this application is: different from the situation of the prior art, according to the technical scheme provided by the application, the sweeping robot obtains the original local map, then determines the point cloud attribute of the occupied grid in the original local map, eliminates the occupied grid of which the point cloud attribute is an outlier, obtains the preprocessed local map, and then performs the operation of extracting the edge starting point on the preprocessed local map to obtain the target edge starting point. Particularly, the method can avoid the influence of interferents on the accuracy and efficiency of the edgewise starting point extraction algorithm by removing the occupied grids with the point cloud attribute of outliers in the original local map, and has a good technical effect.
Drawings
Fig. 1 is a schematic flowchart of an embodiment of an edge starting point extraction method according to the present application;
FIG. 2 is a schematic structural diagram of a local map in an edge starting point extraction method according to the present application;
FIG. 3 is a schematic flow chart illustrating another embodiment of a method for extracting an edge starting point according to the present application;
FIG. 4 is a schematic flow chart illustrating another embodiment of a method for extracting an edge starting point according to the present application;
FIG. 5 is a schematic flow chart illustrating a further embodiment of a method for extracting an edge starting point according to the present application;
FIG. 6 is a schematic flow chart illustrating a further embodiment of a method for extracting an edge starting point according to the present application;
FIG. 7 is a schematic flowchart of a further embodiment of a method for extracting an edge starting point according to the present application;
FIG. 8 is a schematic flow chart illustrating a further embodiment of a method for edge starting point extraction according to the present application;
FIG. 9 is a schematic flowchart of a method for edge starting point extraction according to yet another embodiment of the present application;
fig. 10 is a schematic structural view of an embodiment of a cleaning robot according to the present application;
fig. 11 is a schematic structural diagram of an embodiment of a computer-readable storage medium according to the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not limiting of the application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.
In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless explicitly specifically limited otherwise. Furthermore, the terms "comprising" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by a person skilled in the art that the embodiments described herein can be combined with other embodiments.
Please refer to fig. 1 and fig. 2 simultaneously, in which fig. 1 is a schematic flowchart of an embodiment of an edge starting point extraction method of the present application, and fig. 2 is a schematic structural diagram of a local map of the edge starting point extraction method of the present application. In the current embodiment, the method provided by the present application includes S110 to S130.
S110: the sweeping robot acquires an original local map.
When the sweeping robot executes a sweeping task, an original local map is obtained, the following steps from S120 to S130 are carried out according to the original local map to obtain a target edge starting point, and then the map construction of the home environment is completed based on the determined target edge starting point.
Specifically, in the present embodiment, the local map acquired by the sweeping robot includes occupied grids and unoccupied grids. As illustrated in fig. 2, the acquired original local map is a grid map with the current position of the robot cleaner as the center of the viewpoint (the dot O in fig. 2). The occupied grid in the original local map is identified in fig. 2 as black, the unoccupied grid as white, and the unexplored area as gray.
Further, the sweeping robot can acquire and process the original local map through an acquisition unit arranged by the sweeping robot. In other embodiments, the sweeping robot may also obtain the original local map by acquiring the original local map from other electronic devices dedicated to collecting the original local map.
S120: and determining the point cloud attribute of the occupied grid in the original local map, and removing the occupied grid of which the point cloud attribute is an outlier to obtain the preprocessed local map.
After the original local map is obtained, the point cloud attributes of the occupied grids included in the original local map are further determined respectively. The point cloud attribute of the occupied grid refers to a point cloud type corresponding to the occupied grid in the home environment, and specifically, the point cloud attribute of the occupied grid includes a core point, an outlier and a boundary point.
After determining the point cloud attributes of each occupied grid in the original local map, further removing occupied grids with point cloud attributes being outliers from the original local map, and outputting the local map with all occupied grids with point cloud attributes being outliers as a preprocessed local map for executing the following step S130.
Further, please refer to fig. 3, wherein fig. 3 is a schematic flowchart of another embodiment of an edge starting point extraction method according to the present application. In the present embodiment, the step S120 determines the point cloud attribute of the occupied grid in the original local map, and eliminates the occupied grid whose point cloud attribute is an outlier to obtain the preprocessed local map, which further includes steps S301 to S303.
S301: and respectively counting the number of the first occupation grids included in the first preset area and the second preset area corresponding to the occupation grids, and respectively outputting the first occupation grids and the second occupation grids.
Firstly, respectively counting the number of first occupied grids included in a first preset area and a second preset area corresponding to each occupied data grid aiming at each occupied grid included in an original local map, outputting the number of the first occupied grids included in the first preset area corresponding to each occupied grid as a first number, and outputting the number of the first occupied grids included in the second preset area corresponding to each occupied grid as a second number. In the technical solution provided by the present application, the first occupancy grid is used to refer to other occupancy grids included in the first preset region and the second preset region corresponding to each occupancy grid (i.e., the occupancy grids excluding itself).
For example, in an embodiment, if the original local map includes 10 occupancy grids, in step S301, the first number of the occupancy grids included in the first preset region corresponding to each of the 10 occupancy grids is statistically determined, and the first number of the occupancy grids is output as the first number corresponding to each of the 10 occupancy grids, and the first number of the occupancy grids included in the second preset region corresponding to each of the 10 occupancy grids is determined, and the second number corresponding to each of the occupancy grids is output as the second number corresponding to each of the occupancy grids, so as to be used for determining the point cloud attribute of each of the occupancy grids in the subsequent step.
The first preset area and the second preset area are concentric circle areas taking the occupied grids as circle centers, and the radius of the first preset area is smaller than that of the second preset area. It is understood that in other embodiments, the radius of the first preset area and the radius of the second preset area may be set according to actual requirements.
S302: and respectively determining whether the point cloud attribute occupying the grid is an outlier according to the first quantity and the second quantity corresponding to the occupying grid.
After the number of the first occupied grids included in the first preset area and the second preset area corresponding to the occupied grids is obtained through statistics and is respectively output as the first number and the second number, whether the point cloud attribute of the occupied grids is an outlier is further determined based on the first number and the second number corresponding to the occupied grids.
Further, determining whether the point cloud attribute occupying the grid is an outlier according to the first number and the second number corresponding to the occupying grid respectively, including: and if the first quantity corresponding to the occupied grids is smaller than a first threshold value or the second quantity is smaller than a second threshold value, determining the point cloud attribute of the occupied grids as the outliers. The first threshold and the second threshold are preset empirical values, and the first threshold is smaller than the second threshold.
Specifically, for a certain occupancy grid, if the number of first occupancy grids included in the corresponding first preset region is smaller than a first threshold, or the number of first occupancy grids included in the corresponding second preset region is smaller than a second threshold, the point cloud attribute of the occupancy grid is determined to be an outlier.
Further, if it is determined that the point cloud attribute occupying the grid is not an outlier, the method provided by the application further includes: and further judging whether the point cloud attribute occupying the grid is a core point or a boundary point. Specifically, if the first number corresponding to the occupied grids is greater than the third threshold and the second number is greater than the fourth threshold, the point cloud attribute of the occupied grids is determined to be the core point.
And further determining the point cloud attributes occupying the grids, which are left in the original local map except the point cloud attributes as outliers or core points, as boundary points. In other words, in another embodiment, after determining that the point cloud attribute of the occupied grid is neither an outlier nor a core point, the occupied grid may be automatically classified as a boundary point. It should be noted that in the following description of some embodiments, the occupied grid is referred to by the point cloud attribute of the occupied grid.
In another embodiment, if it is determined that the point cloud attribute occupying the grid is not the outlier after the point cloud attribute occupying the grid is determined to be not the outlier, then the method provided by the present application may also determine the point cloud attribute occupied by the point cloud, that is, the point cloud attribute not occupying the outlier nor the boundary point, as the core point.
It should be noted that, in the above embodiment, it is determined whether the point cloud attribute occupying the grid is an outlier, and then after determining that the point cloud attribute occupying the grid is not an outlier, it is further determined whether the point cloud attribute occupying the grid is a core point and a boundary point in sequence, but in some embodiments, it may also be set to directly determine the point cloud attribute occupying the grid based on a comparison result between the first number corresponding to the occupied grid and the first threshold and the third threshold, and between the second number and the second threshold and the fourth threshold.
S303: if so, eliminating the occupied grids with the point cloud attributes being outliers from the local map, and obtaining the preprocessed local map.
If the point cloud attribute occupying the grid is judged to be the outlier, the occupying grid with the point cloud attribute being the outlier is further removed from the local map. For the occupation grids included in the obtained original local map, respectively judging whether the point cloud attribute of each occupation grid is an outlier or not compared with traversing each occupation grid.
In the current embodiment, before the target edge starting point is determined, whether the occupied grids are outliers or not is determined based on the number of the first occupied grids included in the first preset area and the second preset area corresponding to the occupied grids, and the occupied grids with the point cloud attributes as the outliers are directly removed, so that the interference objects such as human legs, table legs and the like in the home environment are filtered out by removing the outliers, the situation that the interference objects are easily selected as the edge starting point when the interference objects are close to the sweeping robot is avoided, the edge path extracted when the home map is constructed is an invalid path, and the task of constructing the complete map cannot be completed.
S130: and performing edge starting point extraction operation on the preprocessed local map to obtain a target edge starting point.
After the preprocessed local map is obtained, an edge starting point extraction operation is further performed on the obtained preprocessed local map, and then a target edge starting point is obtained. And the target edge starting point is an edge starting point which is finally adopted when the complete home map is constructed.
Further, the edge starting point extracting operation includes: the method comprises a charging pile-based edge starting point extraction step, a straight line feature-based edge starting point extraction step and a point cloud cluster-based edge starting point extraction step. In particular, reference may be made to the following embodiments corresponding to fig. 4 to 9 for specific technical details of the edge starting point extracting operation, which are not described in detail herein.
After the sweeping robot obtains the target edge starting point, a map is further built on the basis of the target edge starting point, so that a complete home map is built.
According to the technical scheme provided in the embodiment corresponding to fig. 1, the sweeping robot obtains the target edge starting point by obtaining an original local map, then determining a point cloud attribute of an occupied grid in the original local map, removing the occupied grid of which the point cloud attribute is an outlier, obtaining a preprocessed local map, and then performing edge starting point extraction operation on the preprocessed local map. Specifically, the influence of an interfering object on the accuracy and efficiency of the edge starting point extraction algorithm can be avoided by removing the occupied grids with the point cloud attribute being outliers in the original local map.
Referring to fig. 4, fig. 4 is a schematic flow chart of another embodiment of an edge starting point extraction method according to the present application. In the present embodiment, the step S130 performs an edge starting point extracting operation on the preprocessed local map, and obtains a target edge starting point, further including steps S401 to S406.
S401: and determining the distance between the current position of the sweeping robot and the position of the charging pile.
After the preprocessed local map is obtained, the distance between the current position of the sweeping robot and the position of the charging pile is further calculated and determined. In particular, it may be provided that the distance between the sweeping robot and the charging pile is determined based on signal interaction time and speed calculation between the two. It is understood that, in other embodiments, the sweeping robot may be configured to calculate and determine the distance between the sweeping robot and the charging pile based on other manners, which are not limited to the above.
S402: and judging whether the distance between the current position of the sweeping robot and the position of the charging pile is smaller than or equal to a first distance threshold value.
After the distance between the current position of the sweeping robot and the position of the charging pile is determined, whether the distance between the current position of the sweeping robot and the position of the charging pile is smaller than or equal to a first distance threshold value or not is further judged. The first distance threshold is a preset distance empirical value used for judging whether the charging pile-based edgewise starting point extraction step can be executed or not, and can be set and adjusted according to actual requirements, and the first distance threshold is not limited by specific numerical values. Further, in other embodiments, the setting of the first distance threshold may be dynamically adjusted according to an actual application condition. For example, when the target edgewise start point still cannot be extracted and obtained based on the following steps (i.e. the target edgewise start point still cannot be obtained through steps S403 to S406), the first distance threshold may be dynamically adjusted further according to the preset value, and then the step S402 is executed again.
If the distance between the current position of the sweeping robot and the position of the charging pile is judged to be smaller than or equal to the first distance threshold value, the following step S403 is executed, and if the distance between the current position of the sweeping robot and the position of the charging pile is judged to be larger than the first distance threshold value, the following step S404 is further executed.
S403: and performing charging pile-based edge starting point extraction on the preprocessed local map to obtain a target edge starting point.
And when the distance between the current position of the sweeping robot and the position of the charging pile is smaller than or equal to the first distance threshold value, further performing an edge starting point extraction step based on the charging pile on the preprocessed local map. Specifically, the technical details of the charging pile-based edgewise starting point extracting step may be referred to in the embodiment corresponding to fig. 5.
S404: and judging whether the straight line feature can be obtained in the preprocessed local map.
Further, the straight line feature is a feature which is relatively ubiquitous in a home scene, and the straight line segment is also suitable to be used as an edge starting point of the sweeper, so that when the distance from the current position of the sweeping robot to the position of the charging pile is judged to be larger than a first distance threshold value, whether the straight line feature can be extracted from the preprocessed local map is further judged.
Specifically, in an embodiment, a hough transform algorithm may be used to extract straight line features in the preprocessed local map, and it may be determined whether the straight line features may be extracted according to an execution result of the hough transform algorithm. If at least one straight line feature can be extracted from the preprocessed local map by the hough transform algorithm, the following step S405 is executed, and if the straight line feature cannot be extracted, the step S406 is further executed.
Further, if in one embodiment, if the extraction of straight line features in the preprocessed local map based on the hough transform algorithm is recorded as
Where i is a number for identifying the extracted straight line feature, and the corresponding L (the largest straight line feature number) can be used to indicate the number of the extracted straight line features,
representing straight line features liThe coordinates of the starting point,
denotes the straight line segment liThe coordinates of the end point. If L is 0, it is determined that the step of extracting the edge starting point based on the straight-line feature cannot be performed on the preprocessed local map, that is, the step S406 needs to be performed at this time.
S405: and executing an edge starting point extraction step based on the straight line features on the preprocessed local map to obtain a target edge starting point.
And if the judgment result shows that at least one straight line feature can be obtained in the preprocessed local map, executing an edge starting point extraction step based on the straight line feature on the preprocessed local map. The step of extracting the starting point along the edge based on the straight line feature may refer to the embodiment corresponding to fig. 6.
S406: and executing an edge starting point extraction step based on point cloud clustering on the preprocessed local map to obtain a target edge starting point.
And if the linear features cannot be obtained in the preprocessed local map, executing an edge starting point extraction step based on point cloud clustering on the preprocessed local map. The step of extracting the starting point along the edge based on the point cloud cluster may refer to an embodiment corresponding to fig. 7.
It should be noted that, in the current embodiment, according to the preset setting, it is determined whether the charging pile-based edge starting point extraction step can be performed on the preprocessed local map, if not, it is determined whether the straight line feature-based edge starting point extraction step can be performed on the preprocessed local map, and when it is determined that the straight line feature-based edge starting point extraction step cannot be performed on the preprocessed local map, the point cloud cluster-based edge starting point extraction step is performed on the preprocessed local map, that is, the point cloud cluster-based edge starting point extraction operation is performed on the preprocessed local map based on the priority order. In other embodiments, the priority order of the extraction steps included in the edge start point extraction operation may also be adjusted based on changes in the home scene or actual application requirements. (i.e., adjusting the execution sequence of the above steps S401 to S406).
Referring to fig. 5, fig. 5 is a schematic flow chart of another embodiment of the method for extracting an edge starting point according to the present application. In the present embodiment, the step S403 executes an edge starting point extraction step based on the charging pile on the preprocessed local map, and obtains a target edge starting point, further including steps S501 to S503.
S501: the charging pile is used as a central point, and the orientation of the charging pile is used as a normal direction to generate a first reference line.
The position of the charging pile is used as a central point, the orientation of the charging pile is used as a normal direction, and a first datum line is generated. The first reference line is used to screen the grid-occupied line meeting the first screening condition in step S502, and the orientation of the charging pile may refer to the orientation of the charging pile on the end face of the sweeping robot.
S502: and screening the occupied grids meeting the first screening condition aiming at the occupied data grids taking each point cloud attribute in the preprocessed local map as a core point.
After the first reference line is generated, an occupation grid which meets a first screening condition and of which the point cloud attribute is the core point is reserved by screening the occupation grid which is included in the preprocessed local map and of which each point cloud attribute is the core point. (in some embodiments, it may also be understood that core points that satisfy the first screening condition are retained).
The first screening condition is that the point cloud attributes are the distance between the occupied grid of the core point and the first reference line and the distance between the occupied grid of the core point and the charging pile, so that the following formula is established. Wherein, the formula is as follows:
wherein λ in the formulad1、λd2、λd3For a predetermined grid-occupied distance threshold, p, for screening point cloud attributes as core pointsBThe generation refers to the position of the charging pile, and a and b are constant parameters of the first datum line in an equation in the preprocessed local map.
S503: and determining one occupying grid which is closest to the charging pile in the occupying grids meeting the first screening condition as a target edge starting point.
After the occupancy grids meeting the first screening condition are screened and obtained, one of the occupancy grids meeting the first screening condition, which is closest to the charging pile, is further determined as a target edge starting point.
In the embodiment illustrated in fig. 5, if the center of the charging pile is taken as a center point, a straight line is generated with the orientation of the charging pile as a normal direction, and an equation of the straight line of the first reference line in the coordinate system of the preprocessed local map is represented as ax + by +1, which is 0.
Occupying grid p of core points for each point cloud attribute retained in the preprocessed local mapiAnd is denoted by pi＝(xi yi). Calculating the distance from the occupied grid of each point cloud attribute to the straight line ax + by +1 which is 0 of the core point and the distance from the occupied grid to the charging pile, further determining whether the occupied grid of the point cloud attribute which is the core point meets the first screening condition or not based on the two calculated distances, and reserving the occupied grid which meets the first screening condition and has the point cloud attribute which is the core point. In other embodiments, the location coordinates of the core point in the preprocessed local map may also be directly substituted into the formula of the first screening condition based on the charging pile and the point cloud attribute, so as to determine whether the core point with the point cloud attribute as the core point meets the first screening condition. And then extracting the occupying grid which is closest to the charging pile from the occupying grids meeting the first screening condition as a target edge starting point.
Referring to fig. 6, fig. 6 is a schematic flowchart illustrating a method for extracting an edge starting point according to another embodiment of the present application. In the present embodiment, the above step S405 performs an edge starting point extraction step based on a straight line feature on the preprocessed local map, and obtains a target edge starting point, further including steps S601 to S604.
S601: a plurality of first linear features are extracted from the preprocessed local map by using a first algorithm.
When the step of extracting the edge starting points of the linear features is executed on the preprocessed local map, firstly, a plurality of first linear features are extracted from the preprocessed local map by using a first algorithm.
Further, the first algorithm comprises a hough transform algorithm. It is understood that in other embodiments, the first algorithm may also include other types of algorithms, not listed here.
S602: and merging and/or screening the plurality of first linear characteristics to obtain at least one second linear characteristic.
After a plurality of first linear features are extracted and obtained from the preprocessed local map, the extracted plurality of first linear features are further merged and/or screened to obtain at least one second linear feature.
It should be noted that, in some embodiments, if only one first linear feature is obtained in the step S601 by extraction, the obtained first linear feature may be further directly output as the second reference line (i.e., step S602 and step S603 are not performed).
Further, merging and/or screening the plurality of first straight line features to obtain at least one second straight line feature, including: aiming at the plurality of first straight line features, merging the first straight line features meeting preset merging conditions, and outputting the merged straight line features and the first straight line features not meeting the preset merging conditions as third straight line features; and screening the straight line features, of which the distance from the sweeping robot is smaller than or equal to a second distance threshold value, as second straight line features for each third straight line feature. The second distance threshold is a preset distance threshold for screening to obtain the second linear feature, and can be set according to actual requirements, and the preset merging condition is that the similarity of the first linear feature is greater than or equal to the preset similarity threshold. Furthermore, the preset merging condition may be set such that an included angle between the different first linear features is smaller than a preset angle threshold, and a distance between the different first linear features is smaller than a preset third distance threshold.
Further, in some embodiments, a second distance threshold may be set dynamically according to the actual extraction condition of the edge starting point. If it is determined that the target edge starting point cannot be obtained based on the initial preset parameters such as the first distance threshold, the second distance threshold and the second threshold, then dynamically adjusting at least one of the preset empirical values according to preset.
In one embodiment, a plurality of first linear features are merged, and then a screening operation is performed on a third linear feature obtained after the merging operation is performed.
In another embodiment, if the merging condition is not met in the first straight-line features, the first straight-line features may be directly output as third straight-line features, and a screening operation may be performed on the third straight-line features.
S603: and determining the second straight line feature with the longest distance as a second reference line.
After obtaining the plurality of second straight line features, the second straight line feature with the longest distance is further determined as the second reference line. And the second reference line is a reference line for screening and obtaining the target edgewise starting point when the step of extracting the edgewise starting point based on the straight line feature is executed.
S604: and determining the point cloud attribute as a core point, the distance from the point cloud attribute to the second reference line to be smaller than a fourth distance threshold value and an occupation grid closest to the sweeping robot as a target edge starting point.
After the second reference line is determined, further determining each point cloud attribute as a distance between the occupied grid of the core point and the second reference line, and determining the point cloud attribute closest to the second reference line as the occupied grid of the core point as the target edgewise starting point. The fourth distance threshold is a preset empirical value.
For the embodiment illustrated in fig. 6, the first linear feature obtained by extraction in the preprocessed partial map by the hough transform algorithm is
And merging the first straight line features with similar features. Specifically, the first straight line features meeting the preset merging condition are merged, specifically, the first straight line features include first straight line features, an included angle between the first straight line features is smaller than a preset angle threshold, and a distance between the first straight line features is smaller than a preset third distance threshold.
Further, if two first straight line features li、ljIf the preset combination condition is met, the four vertex coordinates of the two first linear characteristics are used as input, and the combined result is obtained through least square algorithm fittingThe equation y of the third straight line feature is kx + b, and the combined third straight line feature is expressed as
The coordinates of each endpoint in the combined third straight-line feature may be determined based on the following formula, and the specific formula is as follows:
for the third straight line features output after the merging operation is executed, the distance from the current position of the sweeping robot to each third straight line feature is further calculated, and then the distance greater than a second distance threshold lambda is filteredscopeAnd the distance between the second straight line characteristic and the sweeping robot is smaller than or equal to the second distance threshold lambdascopeIs determined as the second straight line feature.
If a plurality of second straight line features are obtained, further extracting one second straight line feature with the longest length as a second reference line, and further screening an occupation grid with the second reference line distance smaller than a fourth distance threshold value and the point cloud attribute as a core point; and finally, selecting an occupation grid with the point cloud attribute closest to the current sweeping robot position as a core point as a target edge starting point.
Referring to fig. 7, fig. 7 is a schematic flowchart illustrating a method for extracting an edge starting point according to another embodiment of the present application. In the present embodiment, the step S406 performs an edge starting point extraction step based on point cloud clustering on the preprocessed local map, and obtains a target edge starting point, including steps S701 to S704.
S701: and executing the clustering operation of the core points aiming at the occupied grids of which the point cloud attributes included in the preprocessed local map are the core points to obtain a plurality of first point cloud clusters.
Firstly, taking the point cloud attribute contained in the preprocessed local map as an occupied grid of a core point, and taking the Euclidean distance as the measurement and the occupied grid of the point cloud attribute as the core point as the measurementAnd the center is used for clustering a boundary point in the search area with a preset radius and the core point into the same first point cloud cluster (which can also be understood as merging an occupation grid with point cloud attributes as the boundary point into the first point cloud cluster corresponding to the core point). Finally, after traversing the occupation grids with each point cloud attribute as the core point in the preprocessed local map, further obtaining a plurality of first point cloud clusters s1To skAnd further obtaining a first point cloud cluster set S ═ S1 s2 … skWhere k represents the first point cloud cluster number.
It should be noted that the same boundary point may be set to be merged into different first point cloud clusters. In other embodiments, it may also be set that the same boundary point does not appear in different first point cloud clusters, and the boundary point may be clustered to which first point cloud cluster, specifically determined according to time. For example, for the same boundary point, the first point cloud cluster merged first is taken as the criterion, that is, if the same boundary point is merged to the first point cloud cluster corresponding to the other core point before, the boundary point will not be merged to the other first point cloud cluster again.
Further, in an embodiment, after a certain boundary point is merged into a certain first point cloud cluster, a flag may be set for the boundary point, so as to externally inform that the boundary point has been merged into a certain first point cloud cluster.
S702: and merging and/or screening the plurality of first point cloud clusters to obtain at least one second point cloud cluster.
In the current embodiment, after obtaining the plurality of first point cloud clusters, the plurality of first point cloud clusters are merged and/or filtered, so as to obtain at least one second point cloud cluster. Further, whether the merging condition is satisfied may be determined based on the similarity of the different point cloud clusters. In another embodiment, it may be determined whether the current first point cloud cluster and the other point cloud clusters meet a preset point cloud cluster merging condition based on a polar angle value of a core point (the point cloud attribute is an occupied grid of the core point) with a maximum polar angle and/or a minimum polar angle in the first point cloud cluster and a distance between the core point with the maximum polar angle and/or the minimum polar angle and the current position of the robot cleaner.
It should be noted that, in an embodiment, if only one first point cloud cluster is obtained in step S701, the first point cloud cluster may be output as a third point cloud cluster, that is, in this embodiment, steps S702 to S703 are not performed.
In another embodiment, if a plurality of first point cloud clusters are obtained in step S701, step S702 to step S703 are further performed.
Further, please refer to fig. 8, where fig. 8 is a schematic flowchart of a method for extracting an edge starting point according to another embodiment of the present application. In the present embodiment, the step S702 merges and/or filters the plurality of first point cloud clusters to obtain at least one second point cloud cluster, and further includes steps S801 to S804.
S801: and aiming at each first point cloud cluster, sequencing the first point cloud cluster and the preset number of centroid polar angles and the centroid polar angles of the first point cloud cluster to be adjacent to each other according to the numerical value, and determining the first point cloud cluster as a first point cloud cluster subset.
After a plurality of first point cloud clusters are clustered to obtain a plurality of first point cloud clusters, if all the first point cloud clusters are defined as a first point cloud cluster set, further determining, for each obtained first point cloud cluster, a first preset number of other first point cloud clusters adjacent to the first point cloud cluster and the centroid polar angle of the first point cloud cluster in a numerical order as a first point cloud cluster subset.
For example, in an embodiment, S ═ S for the first point cloud cluster set1 s2 … skE.g., if S ═ S 'is obtained in an order from large to small according to the centroid polar angle of each first point cloud cluster'1 s′2 … s′kH, for the first point cloud cluster s'i-1When the preset number is set to be 2, the centroid polar angle and the first point cloud cluster s 'can be sorted according to the preset numerical value'i-1Adjacent first point cloud cluster s'iAnd s'i+1Determining as a first point cloud cluster set s'i-1 s′i s′i+1}。
S802: and aiming at the first point cloud clusters included in the first point cloud cluster subset, judging whether the first point cloud clusters with the maximum and minimum mass center polar angles in the first point cloud cluster subset meet a preset point cloud cluster merging condition or not based on the maximum polar angle and/or the core point polar angle value of the minimum polar angle in the first point cloud clusters and the distance between the core point of the maximum polar angle and/or the minimum polar angle and the sweeping robot.
First, the polar angle is a polar angle in a polar coordinate system established with the center of the sweeping robot as a pole and the traveling direction as a polar axis.
When judging whether at least part of first point cloud clusters in the first point cloud cluster set meet preset point cloud cluster merging conditions, regarding the first point cloud cluster subset { s'i-1 s′i s′i+1Firstly, calculating the current first point cloud cluster s'i-1Polar angle value corresponding to maximum core point of medium polar angle
And the distance from the core point to the current position of the floor sweeping robot
And calculating a first point cloud cluster s'i+1Polar angle value corresponding to minimum core point of medium polar angle
And the distance from the core point to the current position of the sweeping robot
Simultaneously calculating a first point cloud cluster s'iPolar angle values corresponding to maximum and minimum core points of medium polar angle respectively
And
and the first point cloud cluster s'iDistance d from center of mass to current position of sweeping robotiAnd the first point cloudS 'cluster'iNumber n of point clouds included ini. The number of point clouds included in the first point cloud cluster refers to the number of core points and boundary points included in the first point cloud cluster.
After the core point polar angle value of the maximum polar angle and/or the minimum polar angle of the first point cloud cluster included in the first point cloud cluster and the distance between the core point of the maximum polar angle and/or the minimum polar angle and the sweeping robot are respectively determined, the first point cloud cluster s 'is further judged based on the following formula'i-1And s'i+1Whether the point cloud clusters are the same point cloud cluster or not, if so, the first point cloud cluster s'i-1And s'i+1And merging. Wherein, the formula is as follows:
wherein λ isnAnd λθThe method is characterized by comprising the steps of setting a point cloud quantity threshold and an angle threshold in advance.
S803: if yes, merging the first point cloud clusters with the largest and smallest mass center polar angles in the point cloud cluster subset, and outputting the point cloud clusters obtained by merging and the first point cloud clusters which do not meet the point cloud cluster merging conditions as fourth point cloud clusters.
And if the first point cloud clusters with the largest and smallest centroid polar angles in the first point cloud cluster subset meet the preset point cloud cluster merging conditions, merging the two first point cloud clusters, and outputting the point cloud clusters obtained by merging and the first point cloud clusters which do not meet the point cloud cluster merging conditions as a fourth point cloud cluster.
And if the preset point cloud cluster merging condition is not met, sequentially traversing each first point cloud cluster according to the polar angle arrangement sequence of the first point cloud cluster set. If the same first point cloud cluster is judged to meet the preset point cloud cluster merging conditions with a plurality of different first point cloud clusters, merging operation can be performed on the first point cloud cluster and the plurality of different first point cloud clusters meeting the preset point cloud cluster merging conditions.
Further, if it is determined that one or more first point cloud clusters and any one of the first point cloud clusters do not satisfy the preset point cloud cluster merging condition, the first point cloud cluster can be directly output as a fourth point cloud cluster.
After the traversal of each first point cloud cluster included in the first point cloud cluster set is completed, when a plurality of fourth point cloud clusters are obtained, the following step S804 is further performed. In another embodiment, if only one fourth point cloud cluster is obtained after steps S801 to S803, it may be set that the fourth point cloud cluster is directly output as the second point cloud cluster, and step S703 is directly executed or the following step S704 is directly executed.
S804: and determining the distance between the centroid of each fourth point cloud cluster and the sweeping robot, and screening the fourth point cloud clusters with the distance less than or equal to a fifth distance threshold value as second point cloud clusters.
If a plurality of fourth point cloud clusters are obtained, the plurality of fourth point cloud clusters are further screened to obtain a second point cloud cluster based on the distance between the centroid of the fourth point cloud cluster and the floor sweeping robot. Specifically, the distance between the center of mass of each fourth point cloud cluster and the sweeping robot is determined, and the fourth point cloud cluster with the distance smaller than or equal to a fifth distance threshold value is screened as the second point cloud cluster. The fifth distance threshold is a preset distance experience value used for screening the fourth point cloud cluster, and can be specifically set and adjusted according to actual requirements.
S703: and determining the flatness and the flattening core points of the second point cloud cluster, and determining the second point cloud cluster with the highest flatness as a third point cloud cluster.
And after a plurality of second point cloud clusters are obtained, further screening by using the flatness of the second point cloud clusters to obtain a third point cloud cluster. The process of determining the flatness and the flattening core point of the second point cloud cluster may specifically refer to the following embodiment corresponding to fig. 9.
S704: and determining a flat core point in the third point cloud cluster, which is closest to the sweeping robot, as a target edge starting point.
After the third point cloud cluster is determined, a target edgewise starting point is further determined from core flat points included in the third point cloud cluster, and specifically, a flat core point closest to the sweeping robot in the third point cloud cluster is determined as the target edgewise starting point.
Further, please refer to fig. 9, fig. 9 is a schematic flowchart illustrating a method for extracting an edge starting point according to another embodiment of the present application. In the present embodiment, the determining the flatness and the flattening core point of the second point cloud cluster in step S703 further includes steps S901 to S904.
S901: and respectively determining the point clouds in the third preset area for each core point by taking each core point included in the second point cloud cluster as a circle center.
The third preset area is an area determined according to a preset radius. Specifically, each core point included in the second point cloud cluster is respectively used as a circle center, and point clouds in a third preset area are respectively determined for each core point. And the point cloud in the third preset area comprises a core point and a boundary point in the third preset area.
S902: and performing linear fitting on the point cloud in the third preset area by using a second algorithm to obtain a third reference line.
Wherein the second algorithm comprises a least squares algorithm.
Specifically, a straight line fitting may be performed on the retained point cloud by using a least square algorithm to obtain a third reference line. And the third datum line is used for calculating and solving the flatness of the second point cloud cluster.
S903: and determining the average distance from the point cloud included in the third preset area to the third reference line as the flatness of the core point corresponding to the third preset area, and determining the core point with the flatness smaller than the flatness threshold value as the flat core point of the second point cloud cluster.
For each core point (the point cloud attribute is an occupied grid of the core point), the distance from each point cloud included in a third preset area corresponding to each core point to a third reference line is respectively determined, the average distance from each point cloud included in the third preset area to the third reference line is determined, and the average distance is used as the flatness of the core point.
After the flatness of each core point included in the low-hot-point cloud cluster is determined, whether the flatness of the core point is smaller than a flatness threshold value or not is further judged, and if yes, the core points with the flatness smaller than the flatness threshold value are classified as flat core points.
S904: the number of flattened core points included in the second point cloud cluster is determined as the flatness of the second point cloud cluster.
After determining to obtain the flatness of each core point included in the second point cloud cluster, and determining whether each core point is a flattened core point based on the flatness of each core point, further counting the number of flattened core points included in the second point cloud cluster, and then determining the number of flattened core points included in the second point cloud cluster as the flatness of the second point cloud cluster, thereby completing the determination of the flatness and the flattened core points of the second point cloud cluster.
Further, if the target edge starting point still cannot be obtained through the multiple different edge starting point extraction steps in fig. 5 to 7, the value of the second distance threshold and/or the value of the fifth distance threshold may be dynamically adjusted according to the setting, so as to adjust the distance condition for screening the second straight line feature and/or the fourth point cloud cluster, so as to obtain the target edge starting point. It is noted that in some embodiments, the second distance threshold and the fifth distance threshold may be equal.
Referring to fig. 10, fig. 10 is a schematic structural view of an embodiment of a sweeping robot according to the present application. In the present embodiment, the sweeping robot 1000 provided by the present application includes a processor 1001 and a memory 1002 coupled to the processor 1001. The sweeping robot 1000 may perform the method described in any one of the embodiments of fig. 1 to 9 and the corresponding embodiments.
The memory 1002 includes a local storage (not shown) and is used for storing a computer program, and the computer program can implement the method described in any one of the embodiments of fig. 1 to 9 and the corresponding embodiments thereof when executed.
The processor 1001 is coupled to the memory 1002, and the processor 1001 is configured to execute a computer program to perform the method as described in any one of the embodiments of fig. 1 to 9 and corresponding embodiments. Further, in some embodiments, the sweeping robot provided in fig. 10 of the present application can also be extended to include any one of devices that can interact with the sweeping robot, such as a mobile terminal, a computer, an image capturing device with computing storage capability, a server, and the like.
Referring to fig. 11, fig. 11 is a schematic structural diagram of an embodiment of a computer-readable storage medium according to the present application. The computer-readable storage medium 1100 stores a computer program 1101 capable of being executed by a processor, the computer program 1101 being configured to implement the method as described in any one of the embodiments of fig. 1 to 9 and corresponding embodiments thereof. Specifically, the computer-readable storage medium 1100 may be one of a memory, a personal computer, a server, a network device, or a usb disk, and is not limited in any way herein.
The above description is only an embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.
Claims (14)
1. An edge starting point extraction method, the method comprising:
the sweeping robot acquires an original local map;
determining point cloud attributes of occupied grids in the original local map, and removing the point cloud attributes as occupied grids of outliers to obtain a preprocessed local map, wherein the point cloud attributes of the occupied grids comprise core points, the outliers and boundary points;
and executing an edge starting point extraction operation on the preprocessed local map to obtain a target edge starting point.
2. The method of claim 1, wherein determining point cloud attributes of occupancy grids in the original local map and rejecting point cloud attributes as occupancy grids of outliers to obtain a preprocessed local map comprises:
respectively counting the number of first occupied grids included in a first preset area and a second preset area corresponding to the occupied grids, and respectively outputting the first occupied grids and the second occupied grids as a first number and a second number, wherein the first preset area and the second preset area are concentric circle areas with the occupied grids as the circle centers, and the radius of the first preset area is smaller than that of the second preset area;
determining whether the point cloud attribute of the occupied grid is an outlier according to the first quantity and the second quantity corresponding to the occupied grid respectively;
if so, eliminating the occupied grids of the point cloud attributes as the outliers in the local map, and obtaining the preprocessed local map.
3. The method of claim 2, wherein determining whether the point cloud attribute of the occupancy grid is an outlier based on the first and second numbers corresponding to the occupancy grid, respectively, further comprises:
if the first number corresponding to the occupancy grid is smaller than a first threshold or the second number is smaller than a second threshold, determining that the point cloud attribute of the occupancy grid is the outlier, wherein the first threshold is smaller than the second threshold.
4. The method of claim 3, further comprising:
if the first number corresponding to the occupied grid is larger than a third threshold and the second number is larger than a fourth threshold, determining the point cloud attribute of the occupied grid as the core point;
determining point cloud attributes of the occupied grids remaining in the original local map except the point cloud attributes as the outliers or the core points as boundary points.
5. The method according to claim 1, wherein the performing an edge starting point extraction operation on the preprocessed local map to obtain a target edge starting point further comprises:
determining the distance between the current position of the sweeping robot and the position of the charging pile;
and if the distance is smaller than or equal to a first distance threshold value, performing charging pile-based edgewise starting point extraction on the preprocessed local map to obtain the target edgewise starting point.
6. The method of claim 5, wherein the step of charging pile-based edge starting point extraction is performed on the preprocessed local map to obtain the target edge starting point, and further comprising:
generating a first reference line by taking the charging pile as a center point and taking the orientation of the charging pile as a normal direction;
for each point cloud attribute included in the preprocessed local map as an occupancy grid of the core point, screening the occupancy grid satisfying a first screening condition;
and determining one occupying grid which is closest to the charging pile in the occupying grids meeting the first screening condition as the target edgewise starting point.
7. The method of claim 5, wherein if the distance is greater than the first distance threshold, further determining whether a straight-line feature is available in the preprocessed local map;
if yes, performing an edge starting point extraction step based on linear features on the preprocessed local map to obtain the target edge starting point; if not, performing an edge starting point extraction step based on point cloud clustering on the preprocessed local map to obtain the target edge starting point.
8. The method according to claim 7, wherein the step of performing edge starting point extraction based on straight line features on the preprocessed local map to obtain the target edge starting point further comprises:
extracting a plurality of first linear features from the preprocessed local map by using a first algorithm;
merging and/or screening the plurality of first linear features to obtain at least one second linear feature;
determining the second straight line feature with the longest distance as a second reference line;
and determining the point cloud attribute as the core point, the distance between the point cloud attribute and the second reference line is smaller than a fourth distance threshold value, and the occupied grid closest to the sweeping robot as the target edgewise starting point.
9. The method according to claim 8, wherein the merging and/or screening the plurality of first linear features to obtain at least one second linear feature comprises:
aiming at the plurality of first straight line features, merging the first straight line features meeting a preset merging condition, and outputting the merged straight line features and the first straight line features not meeting the preset merging condition as third straight line features;
and screening the straight line feature of which the distance from the sweeping robot is smaller than or equal to a second distance threshold value as a second straight line feature for each third straight line feature.
10. The method according to claim 7, wherein the step of performing point cloud cluster-based edge start extraction on the preprocessed local map to obtain the target edge start comprises:
aiming at the point cloud attribute included in the preprocessed local map and taking the point cloud attribute as the occupation grid of the core point, performing clustering operation of the core point to obtain a plurality of first point cloud clusters;
merging and/or screening the plurality of first point cloud clusters to obtain at least one second point cloud cluster;
determining a flattening core point and a flatness included in the second point cloud cluster, and determining the second point cloud cluster with the highest flatness as a third point cloud cluster;
determining the flattening core point included in the third point cloud cluster, which is closest to the sweeping robot, as the target edge start point.
11. The method of claim 10, wherein the determining a flatness and a flattening core point of the second point cloud cluster comprises:
respectively determining point clouds in a third preset area for each core point by taking each core point included in the second point cloud cluster as a circle center;
performing linear fitting on the point cloud in the third preset area by using a second algorithm to obtain a third reference line;
determining an average distance from the point cloud included in the third preset area to the third reference line as the flatness of the core point corresponding to the third preset area, and determining the core point with the flatness smaller than a flatness threshold value as the flat core point of the second point cloud cluster;
determining the number of the flattening core points included in the second point cloud cluster as the flatness of the second point cloud cluster.
12. The method of claim 10, wherein the merging and/or filtering the plurality of first point cloud clusters to obtain at least one second point cloud cluster, further comprises:
for each first point cloud cluster, sequencing the first point cloud cluster and a preset number of centroid polar angles and the centroid polar angles of the first point cloud cluster into adjacent first point cloud clusters according to the numerical value, and determining the first point cloud clusters as a first point cloud cluster subset;
for the first point cloud clusters included in the first point cloud cluster subset, judging whether the first point cloud clusters with the largest and smallest mass center polar angles in the first point cloud cluster subset meet a preset point cloud cluster merging condition or not based on the core point polar angle value of the largest polar angle and/or the smallest polar angle in the first point cloud clusters and the distance between the core point of the largest polar angle and/or the smallest polar angle and the sweeping robot;
if so, merging the first point cloud clusters with the largest and smallest centroid polar angles in the point cloud cluster subset, and outputting the point cloud clusters obtained by merging and the first point cloud clusters which do not meet the point cloud cluster merging condition as a fourth point cloud cluster;
and/or
And determining the distance between the centroid of each fourth point cloud cluster and the sweeping robot, and screening the fourth point cloud clusters with the distance less than or equal to a fifth distance threshold value as the second point cloud clusters.
13. A sweeping robot, comprising a processor and a memory coupled to the processor; wherein the content of the first and second substances,
the memory is used for storing a computer program;
the processor is configured to run the computer program to perform the method of any one of claims 1 to 12.
14. A computer-readable storage medium, characterized in that it stores a computer program executable by a processor for implementing the method of any one of claims 1 to 12.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111654987.5A CN114529735A (en) | 2021-12-30 | 2021-12-30 | Edge starting point extraction method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111654987.5A CN114529735A (en) | 2021-12-30 | 2021-12-30 | Edge starting point extraction method and related device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114529735A true CN114529735A (en) | 2022-05-24 |
Family
ID=81621085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111654987.5A Pending CN114529735A (en) | 2021-12-30 | 2021-12-30 | Edge starting point extraction method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114529735A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116841300A (en) * | 2023-08-31 | 2023-10-03 | 未岚大陆(北京)科技有限公司 | Working map generation method, working method, control method and related devices |
-
2021
- 2021-12-30 CN CN202111654987.5A patent/CN114529735A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116841300A (en) * | 2023-08-31 | 2023-10-03 | 未岚大陆(北京)科技有限公司 | Working map generation method, working method, control method and related devices |
| CN116841300B (en) * | 2023-08-31 | 2023-12-19 | 未岚大陆(北京)科技有限公司 | Working map generation method, working method, control method and related devices |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020119684A1 (en) | 3d navigation semantic map update method, apparatus and device | |
| CN111190981B (en) | Method and device for constructing three-dimensional semantic map, electronic equipment and storage medium | |
| CN109074084B (en) | Robot control method, device and system and applicable robot | |
| CN100559398C (en) | Automatic deepness image registration method | |
| JP5229744B2 (en) | Image classification device and image classification program | |
| CN108416785B (en) | Topology segmentation method and device for closed space | |
| US9275277B2 (en) | Using a combination of 2D and 3D image data to determine hand features information | |
| CN103970264B (en) | Gesture recognition and control method and device | |
| WO2010135617A1 (en) | Gesture recognition systems and related methods | |
| CN111325204B (en) | Target detection method, target detection device, electronic equipment and storage medium | |
| CN111652974A (en) | Method, device and equipment for constructing three-dimensional face model and storage medium | |
| CN112016469A (en) | Image processing method and device, terminal and readable storage medium | |
| CN108961385B (en) | SLAM composition method and device | |
| CN113658339B (en) | Contour line-based three-dimensional entity generation method and device | |
| CN106778587B (en) | State of flight detection method and device based on unmanned plane image | |
| CN108133516A (en) | A kind of 3D City scenarios automatic reconstruction method | |
| CN106295639A (en) | A kind of virtual reality terminal and the extracting method of target image and device | |
| CN114529735A (en) | Edge starting point extraction method and related device | |
| CN116188683A (en) | Three-dimensional object bounding box determination method, three-dimensional object bounding box determination device, computer equipment and storage medium | |
| US11651533B2 (en) | Method and apparatus for generating a floor plan | |
| CN116721228B (en) | Building elevation extraction method and system based on low-density point cloud | |
| CN111652329B (en) | Image classification method and device, storage medium and electronic equipment | |
| CN115131384B (en) | Bionic robot 3D printing method, device and medium based on edge preservation | |
| CN113326836A (en) | License plate recognition method and device, server and storage medium | |
| CN115775283A (en) | Image compression method, device, storage medium and computer equipment |
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 |