CN115508807A - Point cloud data processing method and device, electronic equipment and storage medium - Google Patents
Point cloud data processing method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115508807A CN115508807A CN202211436085.9A CN202211436085A CN115508807A CN 115508807 A CN115508807 A CN 115508807A CN 202211436085 A CN202211436085 A CN 202211436085A CN 115508807 A CN115508807 A CN 115508807A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- abnormal
- data
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/02—Systems using the reflection of electromagnetic waves other than radio waves
- G01S17/06—Systems determining position data of a target
- G01S17/08—Systems determining position data of a target for measuring distance only
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar, Positioning & Navigation (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Electromagnetism (AREA)
- Multimedia (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The application provides a point cloud data processing method, a point cloud data processing device, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring point cloud data to be processed, wherein the point cloud data to be processed is data generated in laser radar measurement; filtering first point cloud data on a target reference surface in point cloud data to be processed to obtain second point cloud data; clustering the second point cloud data to generate a plurality of point cloud clusters; and detecting preset characteristics of each point cloud cluster to judge whether the point cloud cluster belongs to abnormal point cloud data, wherein the abnormal point cloud data is used for indicating that the point cloud cluster is data with abnormal measuring results when the laser radar is used for measuring. The point cloud on the target reference surface is removed and a clustering algorithm is adopted to generate a point cloud cluster, abnormal point cloud generated by changing a light path through a water drop by a laser beam can be detected according to the preset characteristics of the point cloud cluster, and the influence of the abnormal point cloud on the application of subsequent point cloud is avoided.
Description
Technical Field
The present application relates to the field of laser radar ranging, and in particular, to a method and an apparatus for processing point cloud data, an electronic device, and a storage medium.
Background
When the laser radar is used in a rainy scene, raindrops form water drops on a radar window, when the laser radar performs detection, emitted laser can reach an object to be detected after passing through the water drops, the light path is changed due to the fact that the laser emitted by the laser radar is refracted by the water drops, the laser radar is resolved according to the situation that no water drops are refracted, the error ranging can be generated, and abnormal point clouds can be generated; abnormal point clouds can have serious effects in various application scenarios, such as affecting the normal driving of an autonomous vehicle.
Disclosure of Invention
In order to at least solve the technical problem that abnormal point cloud data generated in rainy days cannot be identified in the related art, the application provides a point cloud data processing method and device, an electronic device and a storage medium.
In a first aspect, the present application provides a method for processing point cloud data, where the method includes:
acquiring point cloud data to be processed, wherein the point cloud data to be processed is data generated in laser radar measurement;
filtering first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data;
generating a plurality of point cloud clusters by clustering the second point cloud data;
and detecting preset characteristics of each point cloud cluster and judging whether the point cloud cluster belongs to abnormal point cloud data, wherein the abnormal point cloud data is used for indicating whether the point cloud cluster is data with abnormal measuring results in laser radar measurement.
Optionally, the acquiring point cloud data to be processed includes:
acquiring original point cloud data;
setting a designated area based on the original point cloud data, and screening out third point cloud data corresponding to the designated area from the original point cloud data;
and taking the third point cloud data as the point cloud data to be processed.
Optionally, the filtering out first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data includes:
determining coordinate position information corresponding to each point in the point cloud data to be processed;
searching the first point cloud data from the point cloud data to be processed according to the coordinate position information corresponding to each point and the position characteristic information corresponding to the point cloud on the target reference surface;
and deleting the first point cloud data in the point cloud data to be processed to obtain the second point cloud data.
Optionally, the detecting the preset feature of each point cloud cluster and determining whether the point cloud data belongs to abnormal point cloud data includes:
taking any one of the point cloud clusters as a target point cloud cluster, and extracting preset characteristics corresponding to the target point cloud cluster; wherein the preset features comprise one or more preset features;
and detecting whether the target point cloud cluster belongs to abnormal point cloud data or not according to the preset characteristics.
Optionally, before detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features, the method further includes:
and screening a sub-feature set from the plurality of preset features according to a pre-training rule, wherein the sub-feature set is an optimal feature combination for detecting whether each point cloud cluster belongs to abnormal point cloud data.
Optionally, the preset feature includes at least one of: the system comprises a gray level feature, a height feature, a point number and a shape feature, wherein the shape feature is determined by the ratio of feature values obtained through principal component analysis.
Optionally, the preset features at least include height features, and the height features are associated with at least one first feature, wherein the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features includes:
aiming at any one first feature in the plurality of first features, determining a first threshold value and a first judgment condition corresponding to the first feature;
comparing a first characteristic value corresponding to the first characteristic with the first threshold value according to the first judgment condition to obtain a first comparison result;
and if the first comparison result shows that the first characteristic value does not meet the first judgment condition, determining that the target point cloud cluster belongs to abnormal point cloud data.
Optionally, the preset features further include a gray feature, and the gray feature is associated with at least one second feature, wherein the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features includes:
if the first comparison result shows that the first feature value does not meet the first judgment condition, determining a second threshold value and a second judgment condition corresponding to any one of the second features;
comparing a second characteristic value corresponding to the two characteristics with the second threshold value according to the second judgment condition to obtain a second comparison result;
and if the second comparison result shows that the second characteristic value does not meet the second judgment condition, determining that the target point cloud cluster belongs to abnormal point cloud data.
Optionally, before the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features, the method further includes:
acquiring a point cloud training sample;
extracting a preset feature sample corresponding to the point cloud training sample, wherein the preset feature sample at least comprises a gray feature sample and a height feature sample;
and training a preset classification model by using at least one first characteristic sample and at least one second characteristic sample corresponding to the gray characteristic sample to search a first threshold value and a first judgment condition corresponding to each first characteristic sample and search a second threshold value and a second judgment condition corresponding to each second characteristic sample.
Optionally, after detecting the preset feature of each point cloud cluster and determining whether the point cloud cluster belongs to abnormal point cloud data, the method further includes:
and if any one of the plurality of point cloud clusters is abnormal point cloud data, deleting the abnormal point cloud data from the point cloud data to be processed.
Optionally, the abnormal point cloud data is abnormal data generated when the lidar window is hung with water.
In a second aspect, the present application provides an apparatus for processing point cloud data, the apparatus comprising:
the device comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring point cloud data to be processed, and the point cloud data to be processed is data generated during laser radar measurement;
the filtering module is used for filtering first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data;
a clustering module for clustering the second point cloud data to generate a plurality of point cloud clusters;
the detection module is used for detecting the preset characteristics of each point cloud cluster and judging whether the point cloud data belong to abnormal point cloud data, wherein the abnormal point cloud data are used for indicating that the point cloud cluster is data with abnormal measuring results when the laser radar is used for measuring.
Optionally, the obtaining module includes:
an acquisition unit for acquiring original point cloud data;
the screening unit is used for setting a designated area based on the original point cloud data and screening third point cloud data corresponding to the designated area from the original point cloud data;
and the first determining unit is used for taking the third point cloud data as the point cloud data to be processed.
Optionally, the filtering module includes:
the second determining unit is used for determining coordinate position information corresponding to each point in the point cloud data to be processed;
the searching unit is used for searching the first point cloud data from the point cloud data to be processed according to the coordinate position information corresponding to each point and the position characteristic information corresponding to the point cloud on the target reference surface;
and the deleting unit is used for deleting the first point cloud data in the point cloud data to be processed to obtain the second point cloud data.
Optionally, the detection module comprises:
the device comprises an extraction unit, a comparison unit and a comparison unit, wherein the extraction unit is used for taking any one of a plurality of point cloud clusters as a target point cloud cluster and extracting preset characteristics corresponding to the target point cloud cluster; wherein the preset features comprise one or more preset features;
and the detection unit is used for detecting whether the target point cloud cluster belongs to abnormal point cloud data or not according to the preset characteristics.
Optionally, before the detecting unit detects whether the target point cloud cluster belongs to abnormal point cloud data according to the preset feature, the apparatus further includes:
and the screening module is used for screening a sub-feature set from the plurality of preset features according to a pre-training rule, wherein the sub-feature set is an optimal feature combination for detecting whether each point cloud cluster belongs to abnormal point cloud data.
Optionally, the preset characteristic includes at least one of: the system comprises a gray level feature, a height feature, a point number and a shape feature, wherein the shape feature is determined by the ratio of feature values obtained through principal component analysis.
Optionally, the preset features at least include a height feature, and the height feature is associated with at least one first feature, wherein the detection unit includes:
a first determining subunit, configured to determine, for any one of the plurality of first features, a first threshold and a first determination condition corresponding to the first feature;
the first comparing subunit is configured to compare, according to the first determining condition, a first feature value corresponding to the first feature with the first threshold value to obtain a first comparing result;
and the second determining subunit is used for determining that the target point cloud cluster belongs to abnormal point cloud data when the first comparison result shows that the first characteristic value does not meet the first judgment condition.
Optionally, the preset feature further includes a grayscale feature, and the grayscale feature is associated with at least one second feature, where the detecting unit includes:
a third determining subunit, configured to determine, when the first comparison result indicates that the first feature value does not satisfy the first determination condition, a second threshold and a second determination condition corresponding to any one of the plurality of second features;
the second comparing subunit is configured to compare a second feature value corresponding to the second feature with the second threshold according to the second determination condition, so as to obtain a second comparison result;
and the fourth determining subunit is configured to determine that the target point cloud cluster belongs to abnormal point cloud data when the second comparison result indicates that the second feature value does not satisfy the second judgment condition.
Optionally, the preset features at least include a grayscale feature and a height feature, the grayscale feature is associated with at least one first feature, and the height feature is associated with at least one second feature, wherein the detecting unit includes:
a first determining subunit, configured to determine, for any one of the plurality of first features, a first threshold and a first determination condition corresponding to the first feature;
the first comparing subunit is configured to compare, according to the first determining condition, a first feature value corresponding to the first feature with the first threshold value to obtain a first comparing result;
a second determining subunit, configured to determine, when the first comparison result indicates that the first feature value does not satisfy the first determination condition, a second threshold and a second determination condition corresponding to any one of the plurality of second features;
the second comparing subunit is configured to compare, according to the second determining condition, a second feature value corresponding to the second feature with the second threshold value to obtain a second comparison result;
and the third determining subunit is configured to determine that the target point cloud cluster belongs to abnormal point cloud data when the second comparison result indicates that the second feature value does not satisfy the second judgment condition.
Optionally, before the detecting unit detects whether the target point cloud cluster belongs to abnormal point cloud data according to the preset feature, the apparatus further includes:
the second acquisition module is used for acquiring a point cloud training sample;
the extraction module is used for extracting a preset feature sample corresponding to the point cloud training sample, wherein the preset feature sample at least comprises a gray feature sample and a height feature sample;
the searching module is used for training a preset classification model by using at least one first characteristic sample and at least one second characteristic sample corresponding to the gray characteristic sample so as to search a first threshold value and a first judging condition corresponding to each first characteristic sample and search a second threshold value and a second judging condition corresponding to each second characteristic sample.
Optionally, after the detecting module detects the preset feature of each point cloud cluster and determines whether the point cloud cluster belongs to abnormal point cloud data, the apparatus further includes:
and the deleting module is used for deleting the abnormal point cloud data from the point cloud data to be processed when any one of the plurality of point cloud clusters is abnormal point cloud data.
Optionally, the abnormal point cloud data is abnormal data generated when the laser radar window is hung with water.
In a third aspect, the present application provides a vehicle, which detects abnormal point cloud data in data generated during laser radar measurement according to the point cloud data processing method of any one of the first aspect.
In a fourth aspect, the present application provides an electronic device, comprising: a processor and a memory, wherein the memory is configured to store computer readable instructions; a processor, coupled to the memory, for implementing the method steps of the first aspect when executing the computer-readable instructions.
In a fifth aspect, the present application provides a computer readable storage medium having stored thereon computer readable instructions which, when executed by a processor, implement the method steps of the first aspect described above.
In a sixth aspect, embodiments of the present application further provide a computer program product containing instructions, which when executed on a computer, cause the computer to perform the method steps of the first aspect.
Compared with the related art, the technical scheme provided by the embodiment of the application has the following advantages:
according to the point cloud data processing method provided by the embodiment of the application, point cloud data (namely first point cloud data) on a target reference surface in the point cloud data to be processed is filtered aiming at the point cloud data to be processed of the laser radar; then, clustering the residual point cloud (namely second point cloud data) of the point cloud removed from the target reference surface to obtain a plurality of point cloud clusters; and finally, detecting the preset characteristics of each point cloud cluster and judging whether the point cloud cluster belongs to abnormal point cloud data. According to the embodiment of the invention, the point cloud on the target reference surface is removed and a clustering algorithm is adopted to generate the point cloud cluster, and the abnormal point cloud generated by the laser beam changing the light path through the water drop can be detected according to the preset characteristics of the point cloud cluster, so that the influence of the abnormal point cloud on the application of the subsequent point cloud is avoided.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and together with the description, serve to explain the principles of the application.
In order to more clearly illustrate the embodiments of the present application 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 described below, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is an example application provided in accordance with the present application;
FIG. 2 is a schematic diagram of raw point cloud data provided in accordance with the present application;
fig. 3 is an application scenario of a point cloud data processing method according to an embodiment of the present application;
fig. 4 is a schematic flowchart of a method for processing point cloud data according to an embodiment of the present disclosure;
fig. 5 is a schematic flowchart of another method for processing point cloud data according to an embodiment of the present disclosure;
fig. 6 is a judgment model structure for abnormal point cloud detection according to an embodiment of the present disclosure;
fig. 7 is a schematic flowchart of an abnormal point cloud processing method according to an embodiment of the present application;
FIG. 8 is a schematic diagram illustrating the abnormal point cloud data being removed according to the present application;
fig. 9 is a schematic structural diagram of a device for processing point cloud data according to an embodiment of the present disclosure;
fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, 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, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In general, lidar suffers from two main problems when used in rainy weather scenarios: one is aerial raindrop, when the laser of radar transmission hit the raindrop on, can produce corresponding range finding, the noise point that this kind of aerial raindrop produced is discrete degree bigger usually, because can filter through methods such as going the outlier, another problem is that the raindrop falls on the window of radar, form the drop on the radar window, this kind of condition can lead to the laser of radar transmission to pass through the refraction of drop, lead to the energy light path to change, produce unusual some clouds, this kind of some clouds can't adopt the mode of going the outlier to get rid of. As shown in fig. 1, fig. 1 shows an example application provided by the present application, when an ideal light ray passes through a water bead refraction angle upward, the actual point cloud ranging is lengthened, i.e. a point depressed from the ground is generated, and when the ideal light ray passes through the water bead refraction angle downward, the actual point cloud ranging is shortened, i.e. a point raised from the ground is generated. Therefore, when the window is hung with water, the originally planar point cloud generates convex and concave point clouds similar to obstacles, as shown in the actual point cloud of fig. 2, the white point cloud is abnormal, the actual scene is the ground, but the point cloud is protruded and concave on the ground due to refraction generated by hanging water, so that the detection algorithm considers that the position has one obstacle, and normal running of the automatic driving vehicle is influenced.
In order to solve the above problems, the present application provides a method for processing point cloud data, which may be applied to a laser radar, or may be applied to a hardware environment formed by a laser radar and a server. The server is connected with the server through the network the laser radar makes the connection. The network may be a wide area network such as a 4G or 5G network.
First, referring to fig. 3, fig. 3 is an application scenario of a point cloud data processing method provided in an embodiment of the present application, and first, referring to fig. 3, it shows an application scenario in which an embodiment of the present invention may be implemented. The scenario shown in fig. 3 includes a laser radar system 200 and a server 100 in a vehicle. Lidar system 200 uses infrared light or laser light to measure the position, path, etc. of the vehicle, acting as the eyes of the vehicle, allowing the vehicle to always look "in" all directions. The server 100 may be a software operator server, a cloud server, or the like. The laser radar system 200 and the server 100 may be connected to each other by communication via any of the networks.
The point cloud data processing method according to the embodiment of the present application may be executed by the server 100, or may be executed by the laser radar system 200, or may be executed by both the server 100 and the laser radar system 200.
Fig. 4 is a schematic flow chart of a method for processing point cloud data according to an embodiment of the present disclosure. As shown in fig. 4, the method comprises the steps of:
step S402, point cloud data to be processed is obtained, wherein the point cloud data to be processed is data generated in laser radar measurement;
the point cloud data to be processed in the embodiment is data generated during laser radar measurement; further, the point cloud is data of the laser radar ranging and visual display of angle results.
In a specific implementation, the step S402 specifically includes the following steps:
step S501, acquiring original point cloud data;
step S502, setting a designated area based on the original point cloud data, and screening out third point cloud data corresponding to the designated area from the original point cloud data;
after the laser radar scans a scene to be detected, a designated area in the scene can be set, and the point cloud data of the designated area is detected, namely the point cloud data of an ROI (region of interest) is set.
And S503, taking the third point cloud data as point cloud data to be processed.
For example, a certain ROI is set for the input point cloud, and the point cloud in the ROI is selected for detection, where the point cloud in the ROI may be a range in which an abnormal point cloud is actually easily generated, and the range selection is determined by a range required to affect a subsequent planning algorithm.
Step S404, filtering first point cloud data on a target reference surface in point cloud data to be processed to obtain second point cloud data;
the point cloud obtained by scanning the laser radar contains a large number of surface points (namely the target reference surface), which brings troubles to the operations of clustering, identifying and the like of subsequent point clouds, so that the subsequent point clouds can be filtered firstly, on one hand, the data volume can be reduced, and on the other hand, the interference to the subsequent operations is avoided.
In specific implementation, the step S404 includes the following steps:
step S601, determining coordinate position information corresponding to each point in point cloud data to be processed;
in the embodiment, the position coordinates of each point in the point cloud data to be processed obtained by scanning are known;
step S602, searching first point cloud data in point cloud data to be processed according to coordinate position information corresponding to each point and position characteristic information corresponding to point cloud on a target reference surface;
and searching first point cloud data in the point cloud to be processed according to the position coordinates of each point and the set position characteristics of the point cloud on the target reference surface.
The first point cloud data is screened out by using the position characteristics of each point position and the point cloud on the target reference surface, and the method comprises the following steps: horizontal plane calibration, normal vector, grid height difference, absolute height, average height, etc. Thereby finding out the point cloud on the real target reference surface in the point cloud to be processed.
Step S603, delete the first point cloud data in the point cloud data to be processed, and obtain the second point cloud data.
And deleting the point cloud on the target reference surface in the point cloud to be processed so as to prepare data by adopting a machine learning method subsequently.
Step S406, clustering the second point cloud data to generate a plurality of point cloud clusters;
in this embodiment, the second point cloud data, that is, the remaining point cloud data of the point cloud on the target reference surface filtered out from the point cloud data to be processed is processed by using a clustering algorithm to obtain a plurality of point cloud clusters.
The clustering algorithm, also called group analysis, is a statistical analysis method for studying (sample or index) classification problems. The clustering algorithm in this embodiment includes, but is not limited to, the following: european cluster, DBSCAN, etc.
For example, a point cloud that is relatively close in distance may be treated as a cluster based on the known three-dimensional coordinates of the various points.
Step S408, detecting the preset characteristics of each point cloud cluster and judging whether the point cloud cluster belongs to abnormal point cloud data, wherein the abnormal point cloud data is used for indicating that the point cloud cluster is data with abnormal measuring results when the laser radar is used for measuring.
The method comprises the steps of obtaining point cloud data to be processed, and filtering the point cloud data on a target reference surface in the point cloud data to be processed; then, clustering the residual point clouds (namely second point cloud data) of the point clouds on the removed target reference surface to obtain a plurality of point cloud clusters; and finally, detecting the preset characteristics of each point cloud cluster and judging whether the point cloud cluster belongs to abnormal point cloud data. According to the embodiment of the invention, the point cloud on the target reference surface is removed and a clustering algorithm is adopted to generate the point cloud cluster, and the abnormal point cloud generated by changing the light path of the laser beam through the water drop can be detected according to the preset characteristics of the point cloud cluster, so that the influence of the abnormal point cloud on the application of the subsequent point cloud is avoided.
In an optional embodiment of the present disclosure, the method further includes: taking any one of the point cloud clusters as a target point cloud cluster, and extracting preset features corresponding to the target point cloud cluster, wherein the preset features comprise one or more preset features; and detecting whether the target point cloud cluster belongs to abnormal point cloud data or not according to preset characteristics.
Further, in order to improve the accuracy of detecting whether the point cloud belongs to the abnormal point cloud data by using the preset features, before detecting whether the target point cloud cluster belongs to the abnormal point cloud data according to the preset features, the method further comprises the following steps: and screening out a sub-feature set from a plurality of preset features according to a pre-training rule, wherein the sub-feature set is an optimal feature combination for detecting whether each point cloud cluster belongs to abnormal point cloud data.
Through the embodiment, at least one preset feature is extracted from each point cloud, an optimal feature combination for detecting abnormal point clouds is screened out from a plurality of preset features according to a feature combination (namely the pre-training rule) learned through pre-training, point cloud data is detected through the optimal feature combination, for example, any one of height features, gray scale features, points and shape features can be used for detection, any two features and any three features can be used for detection, and the accuracy of detection is improved.
Preferably, the preset features in the present solution include at least one of: the system comprises a gray level feature, a height feature, a point number and a shape feature, wherein the shape feature is determined by the ratio of feature values obtained through principal component analysis.
In this embodiment, the gray scale and the position of each point in the point cloud data to be processed are known, and the gray scale features corresponding to each point cloud cluster are extracted, so that the maximum gray scale value, the minimum gray scale value, the average gray scale value, the gray scale variance and the like can be obtained; extracting the height characteristics corresponding to each point cloud cluster, obtaining the maximum height value, the minimum height value, the average height value, the variance and the point number of the point cloud cluster relative to the ground, analyzing the characteristic vector, the characteristic value and the proportion of the point cloud cluster through principal component analysis, calculating the covariance matrix of the point cloud data matrix, then obtaining the characteristic value and the characteristic vector of the covariance matrix, selecting the k characteristics with the maximum characteristic value (namely the maximum variance) and the proportion thereof, and the matrix formed by the corresponding characteristic vectors, obtaining the shape corresponding to each point cloud cluster, and judging whether each point cloud cluster is abnormal based on whether the shape of the point cloud cluster is abnormal.
In an optional embodiment of the present disclosure, the preset features at least include height features, and the height features are associated with at least one first feature, wherein detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features includes the following steps:
step S701, aiming at any one of a plurality of first features, determining a first threshold value and a first judgment condition corresponding to the first feature;
in the present embodiment, the height features are associated with at least a height maximum value, a height minimum value, a height average value, and the like (i.e., first features); the different first characteristics correspond to different judgment conditions. For example, the height maximum value corresponds to one determination condition, the height minimum value corresponds to one determination condition, and the like.
Step S702, comparing a first characteristic value corresponding to the first characteristic with a first threshold value according to a first judgment condition to obtain a first comparison result;
step S703, if the first comparison result indicates that the first feature value does not satisfy the first determination condition, it is determined that the target point cloud cluster belongs to abnormal point cloud data.
In this embodiment, one height feature may be used to perform abnormal point cloud detection on a target point cloud cluster.
Referring to fig. 6, fig. 6 is a schematic flow chart of an abnormal point cloud processing method according to an embodiment of the present disclosure. For example, the height maximum is compared to a threshold of-0.428 to determine if the height maximum is less than or equal to-0.428.
As shown in fig. 6, if the maximum height value Z _ max (i.e. the first feature) is greater than or equal to-0.428 (i.e. the first threshold), the target point cloud cluster is a normal point cloud; otherwise, step S704 is executed to determine a second threshold and a second determination condition corresponding to any second feature associated with the grayscale feature, and step S705 is executed.
Step S705, comparing a second characteristic value corresponding to the second characteristic with a second threshold value according to a second judgment condition to obtain a second comparison result;
step S706, if the second comparison result shows that the second characteristic value does not meet the second judgment condition, determining that the target point cloud cluster belongs to abnormal point cloud data.
In this embodiment, the adopted optimal feature combination is the height feature and the grayscale feature, and under the condition that any one of the first features of the height feature does not satisfy the corresponding first judgment condition, any one of the second features in the grayscale feature is continuously judged, whether the second judgment condition is satisfied or not, so that the accuracy of point cloud detection is improved.
In a specific implementation process, it may be determined whether the maximum intensity _ max of the gray scale (i.e., the second feature) is less than or equal to 6.5 (i.e., the second threshold); if intensity _ max > =6.5, the target point cloud cluster is a normal point cloud; otherwise, the target point cloud cluster is abnormal point cloud.
Further, before detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the gray feature and the height feature, the method further comprises the following steps:
step S801, acquiring a point cloud training sample;
step S802, extracting a preset feature sample corresponding to the point cloud training sample, wherein the preset feature sample at least comprises a gray feature sample and a height feature sample;
step S803, training a preset classification model by using at least one first feature sample and at least one second feature sample corresponding to the grayscale feature sample, so as to find out a first threshold and a first judgment condition corresponding to each first feature sample, and find out a second threshold and a second judgment condition corresponding to each second feature sample.
In this embodiment, the classification model may be a decision tree model or other classification models capable of learning the optimal features and the determination threshold, which is not limited herein.
After step S408, the method further includes: and when any one of the plurality of point cloud clusters is abnormal point cloud data, deleting the abnormal point cloud data from the point cloud data to be processed. Through the embodiment, if the abnormal point cloud exists, the abnormal point cloud is removed from the point cloud to be processed, and finally data with the abnormal point cloud removed are obtained.
According to the scheme, abnormal point cloud data generated when the laser radar window is hung with water can be effectively detected, so that the problem caused by the abnormal point cloud is solved, and the laser radar can be normally used in rainy days.
By way of example, as shown in fig. 7, the following steps are included:
step S901, inputting an original point cloud;
step S902, setting a certain ROI;
the method comprises the steps of setting a certain ROI (region of interest) for an input point cloud, selecting the point cloud in the ROI for further processing, wherein the ROI can generally take the point cloud in a certain distance range, and the range is determined by the range which is easy to generate abnormal point cloud actually and the range which is required by the subsequent planning control algorithm.
In step S903, ground points (i.e., points on the target reference plane) are removed by a ground point detection algorithm.
And step S904, clustering the rest point clouds.
The point clouds with the ground points removed remaining are clustered in this example.
Step S905, extracting features of the point cloud of each class.
Extracting characteristics, such as maximum, minimum, average, variance and the like, of the gray level intensity of each clustered point cloud; the height z value is maximum, minimum, average, variance, point number, principal component analysis eigenvector, eigenvalue and proportion.
And step S906, judging whether the abnormal point cloud exists or not according to the characteristics, and deleting the abnormal point cloud if the abnormal point cloud exists.
In this embodiment, the method includes two parts, namely model training and actual deployment:
and 1, during model training, taking the characteristics of each clustered point cloud as input, and taking whether the point cloud is abnormal in water hanging as output.
For example, the optimal features and the judgment threshold are obtained through a decision tree, the type of the point cloud of the cluster is obtained by only calculating the corresponding features and judging when the point cloud is deployed, and if the point cloud is abnormal, the point cloud is deleted to obtain the point cloud with the abnormal water hanging removed finally.
For example, a final judgment model obtained through decision tree analysis is shown in fig. 6, a node of each tree corresponds to a judgment condition and a threshold value and a corresponding class, the model includes two categories, normal is a normal point cloud, error represents an abnormal point cloud generated by window hanging water, that is, the maximum value of the height z value of the midpoint of each cluster of point cloud is < = -0.428, and the maximum value of the gray level intensity is < =6.5 is an abnormal point cloud generated by window hanging water.
2, in the deployment part, corresponding features are calculated and the radar of the point cloud is obtained through judgment for each point cloud according to the optimal judgment condition and the threshold value obtained in the training process, and if the point cloud is abnormal, the point cloud is removed from the original point cloud, and finally data of the point cloud without the abnormal point is obtained. As shown in fig. 8, the white point cloud is an abnormal point cloud detected due to the hanging of water from the radar window, and the abnormal point cloud can be effectively detected by the method.
The method comprises the steps of clustering obstacles to be processed by a ground point removing and clustering algorithm for abnormal point clouds generated by window hanging water, extracting characteristics of the point clouds of each obstacle, and classifying and identifying the abnormal point clouds through a classification model, so that the abnormal point clouds generated by laser radar hanging water are prevented from being taken as the obstacles to influence the normal running of an automatic driving vehicle.
Based on the method for processing point cloud data provided in the foregoing embodiments, based on the same inventive concept, a device for processing point cloud data is also provided in this embodiment, and the device is used to implement the foregoing embodiments and preferred embodiments, and is not described again after having been described. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 9 is a schematic structural diagram of a device for processing point cloud data according to an embodiment of the present application, as shown in fig. 9, the device includes:
the first acquisition module 90 is configured to acquire point cloud data to be processed, where the point cloud data to be processed is data generated during laser radar measurement;
a filtering module 92, connected to the obtaining module 90, for filtering first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data;
a clustering module 94, connected to the filtering module 92, for clustering the second point cloud data to generate a plurality of point cloud clusters;
and the detection module 96 is connected to the clustering module 94 and is configured to detect preset features of each point cloud cluster and determine whether the point cloud data belongs to abnormal point cloud data, where the abnormal point cloud data is used to indicate that the point cloud cluster is data with abnormal measurement results when the laser radar measures distance.
Optionally, the obtaining module 90 includes:
an acquisition unit for acquiring original point cloud data;
the screening unit is used for setting a specified area based on the original point cloud data and screening out third point cloud data corresponding to the specified area from the original point cloud data;
and the first determining unit is used for taking the third point cloud data as point cloud data to be processed.
Optionally, the filtering module 92 comprises:
the second determining unit is used for determining coordinate position information corresponding to each point in the point cloud data to be processed;
the searching unit is used for searching first point cloud data from the point cloud data to be processed according to the coordinate position information corresponding to each point and the position characteristic information corresponding to the point cloud on the target reference surface;
and the deleting unit is used for deleting the first point cloud data in the point cloud data to be processed to obtain the second point cloud data.
Optionally, the detection module 96 comprises:
the extraction unit is used for taking any one of the point cloud clusters as a target point cloud cluster and extracting preset characteristics corresponding to the target point cloud cluster; wherein the preset features comprise one or more preset features;
and the detection unit is used for detecting whether the target point cloud cluster belongs to abnormal point cloud data or not according to the preset characteristics.
Optionally, before the detecting unit detects whether the target point cloud cluster belongs to abnormal point cloud data according to the preset feature, the apparatus further includes:
and the screening module is used for screening a sub-feature set from a plurality of preset features according to a pre-training rule, wherein the sub-feature set is an optimal feature combination for detecting whether each point cloud cluster belongs to abnormal point cloud data.
Optionally, the preset features include at least one of: the system comprises a gray level feature, a height feature, a point number and a shape feature, wherein the shape feature is determined by the ratio of feature values obtained through principal component analysis.
Optionally, the preset features at least include a height feature, and the height feature is associated with at least one first feature, wherein the detection unit includes:
the first determining subunit is configured to determine, for any one of the plurality of first features, a first threshold and a first determination condition corresponding to the first feature;
the first comparison subunit is configured to compare a first feature value corresponding to the first feature with a first threshold according to a first determination condition, so as to obtain a first comparison result;
and the second determining subunit is used for determining that the target point cloud cluster belongs to abnormal point cloud data when the first comparison result shows that the first characteristic value does not meet the first judgment condition.
Optionally, the preset features further include a grayscale feature, and the grayscale feature is associated with at least one second feature, wherein the detection unit includes:
a third determining subunit, configured to determine, when the first comparison result indicates that the first feature value does not satisfy the first determination condition, a second threshold and a second determination condition corresponding to the second feature for any second feature of the plurality of second features;
the second comparison subunit is used for comparing a second characteristic value corresponding to the second characteristic with a second threshold value according to a second judgment condition to obtain a second comparison result;
and the fourth determining subunit is used for determining that the target point cloud cluster belongs to abnormal point cloud data when the second comparison result shows that the second characteristic value does not meet the second judgment condition.
Optionally, before the detecting unit detects whether the target point cloud cluster belongs to the abnormal point cloud data according to the preset feature, the apparatus further includes:
the second acquisition module is used for acquiring a point cloud training sample;
the extraction module is used for extracting preset feature samples corresponding to the point cloud training samples, wherein the preset feature samples at least comprise gray feature samples and height feature samples;
the searching module is used for training a preset classification model by using at least one first characteristic sample and at least one second characteristic sample corresponding to the gray characteristic sample so as to search a first threshold value and a first judging condition corresponding to each first characteristic sample and search a second threshold value and a second judging condition corresponding to each second characteristic sample.
Optionally, after the detecting module 96 detects the preset feature of each point cloud cluster to determine whether the point cloud cluster belongs to the abnormal point cloud data, the apparatus further includes:
and the deleting module is used for deleting the abnormal point cloud data from the point cloud data to be processed when any one of the plurality of point cloud clusters is abnormal point cloud data.
Optionally, the abnormal point cloud data is abnormal data generated when the lidar window is hung with water.
It should be noted that, the above modules may be implemented by software or hardware, and for the latter, the following may be implemented, but not limited to: the modules are all positioned in the same processor; alternatively, the modules are respectively located in different processors in any combination.
As shown in fig. 10, an electronic device provided in the embodiment of the present application includes a processor 111 and a memory 113, where,
a memory 113 for storing computer readable instructions;
in an embodiment of the present application, the processor 111, connected to the memory 113, is configured to implement the method for processing point cloud data provided in any one of the method embodiments described above when executing the readable instructions stored on the memory 113.
The present application further provides a computer-readable storage medium, on which computer-readable instructions are stored, and when executed by a processor, the computer-readable instructions implement the steps of the method for processing point cloud data provided in any one of the foregoing method embodiments.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The above description is merely illustrative of particular embodiments of the invention that enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (15)
1. A method for processing point cloud data, the method comprising:
acquiring point cloud data to be processed, wherein the point cloud data to be processed is data generated in laser radar measurement;
filtering first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data;
generating a plurality of point cloud clusters by clustering the second point cloud data;
and detecting preset characteristics of each point cloud cluster to judge whether the point cloud cluster belongs to abnormal point cloud data, wherein the abnormal point cloud data is used for indicating that the point cloud cluster is data with abnormal measuring results when the laser radar is used for measuring.
2. The method of claim 1, wherein the obtaining point cloud data to be processed comprises:
acquiring original point cloud data;
setting a designated area based on the original point cloud data, and screening out third point cloud data corresponding to the designated area from the original point cloud data;
and taking the third point cloud data as the point cloud data to be processed.
3. The method of claim 1, wherein the filtering out the first point cloud data on the target reference surface from the point cloud data to be processed to obtain the second point cloud data comprises:
determining coordinate position information corresponding to each point in the point cloud data to be processed;
searching the first point cloud data from the point cloud data to be processed according to the coordinate position information corresponding to each point and the position characteristic information corresponding to the point cloud on the target reference surface;
and deleting the first point cloud data in the point cloud data to be processed to obtain the second point cloud data.
4. The method of claim 1, wherein the detecting the preset features of each point cloud cluster to determine whether the point cloud data belongs to abnormal point cloud data comprises:
taking any one of the point cloud clusters as a target point cloud cluster, and extracting preset characteristics corresponding to the target point cloud cluster; wherein the preset features comprise one or more preset features;
and detecting whether the target point cloud cluster belongs to abnormal point cloud data or not according to the preset characteristics.
5. The method of claim 4, wherein prior to detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features, the method further comprises:
and screening a sub-feature set from the plurality of preset features according to a pre-training rule, wherein the sub-feature set is an optimal feature combination for detecting whether each point cloud cluster belongs to abnormal point cloud data.
6. The method of claim 4, wherein the predetermined characteristic comprises at least one of: the method comprises the following steps of gray level characteristics, height characteristics, point numbers and shape characteristics, wherein the shape characteristics are determined by the ratio of characteristic values obtained through principal component analysis.
7. The method according to any one of claims 6, wherein the preset features at least comprise height features, the height features are associated with at least one first feature, and wherein the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features comprises:
aiming at any one first feature in the plurality of first features, determining a first threshold value and a first judgment condition corresponding to the first feature;
comparing a first characteristic value corresponding to the first characteristic with the first threshold value according to the first judgment condition to obtain a first comparison result;
and if the first comparison result shows that the first characteristic value does not meet the first judgment condition, determining that the target point cloud cluster belongs to abnormal point cloud data.
8. The method of claim 7, wherein the predetermined features further comprise a gray scale feature associated with at least a second feature, wherein the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the predetermined features comprises:
if the first comparison result shows that the first feature value does not meet the first judgment condition, determining a second threshold value and a second judgment condition corresponding to any one of the second features;
comparing a second characteristic value corresponding to the two characteristics with the second threshold value according to the second judgment condition to obtain a second comparison result;
and if the second comparison result shows that the second characteristic value does not meet the second judgment condition, determining that the target point cloud cluster belongs to abnormal point cloud data.
9. The method of claim 8, wherein prior to the detecting whether the target point cloud cluster belongs to abnormal point cloud data according to the preset features, the method further comprises:
acquiring a point cloud training sample;
extracting a preset feature sample corresponding to the point cloud training sample, wherein the preset feature sample at least comprises a gray feature sample and a height feature sample;
and training a preset classification model by using at least one first characteristic sample and at least one second characteristic sample corresponding to the gray characteristic sample to search a first threshold value and a first judgment condition corresponding to each first characteristic sample and search a second threshold value and a second judgment condition corresponding to each second characteristic sample.
10. The method of claim 1, wherein after detecting the preset features of each point cloud cluster to determine whether the point cloud data belongs to abnormal point cloud data, the method further comprises:
and if any one of the plurality of point cloud clusters is abnormal point cloud data, deleting the abnormal point cloud data from the point cloud data to be processed.
11. The method of any one of claims 1-10, wherein the anomalous point cloud data is anomalous data generated when a lidar window is flooded.
12. An apparatus for processing point cloud data, the apparatus comprising:
the device comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring point cloud data to be processed, and the point cloud data to be processed is data generated during laser radar measurement;
the filtering module is used for filtering first point cloud data on a target reference surface in the point cloud data to be processed to obtain second point cloud data;
a clustering module for clustering the second point cloud data to generate a plurality of point cloud clusters;
the detection module is used for detecting the preset characteristics of each point cloud cluster and judging whether the point cloud cluster belongs to abnormal point cloud data, wherein the abnormal point cloud data is used for indicating that the point cloud cluster is data with abnormal measurement results when the laser radar is used for measuring.
13. A vehicle characterized in that abnormal point cloud data in data generated at the time of laser radar measurement is detected according to the method for processing point cloud data of any one of claims 1 to 11.
14. An electronic device, comprising: a processor and a memory;
the memory for storing computer readable instructions;
the processor, coupled to the memory, is configured to implement the method steps of any of claims 1-11 when executing the computer-readable instructions.
15. A computer-readable storage medium having computer-readable instructions stored therein, which when executed by a processor, perform the method steps of any one of claims 1-11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211436085.9A CN115508807B (en) | 2022-11-16 | 2022-11-16 | Point cloud data processing method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211436085.9A CN115508807B (en) | 2022-11-16 | 2022-11-16 | Point cloud data processing method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115508807A true CN115508807A (en) | 2022-12-23 |
| CN115508807B CN115508807B (en) | 2023-06-30 |
Family
ID=84514011
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211436085.9A Active CN115508807B (en) | 2022-11-16 | 2022-11-16 | Point cloud data processing method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115508807B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108052624A (en) * | 2017-12-15 | 2018-05-18 | 深圳市易成自动驾驶技术有限公司 | Processing Method of Point-clouds, device and computer readable storage medium |
| US10634793B1 (en) * | 2018-12-24 | 2020-04-28 | Automotive Research & Testing Center | Lidar detection device of detecting close-distance obstacle and method thereof |
| CN112508912A (en) * | 2020-12-07 | 2021-03-16 | 中联重科股份有限公司 | Ground point cloud data filtering method and device and boom anti-collision method and system |
| WO2021097618A1 (en) * | 2019-11-18 | 2021-05-27 | 深圳市大疆创新科技有限公司 | Point cloud segmentation method and system, and computer storage medium |
| CN114219770A (en) * | 2021-11-26 | 2022-03-22 | 深圳市优必选科技股份有限公司 | Ground detection method, ground detection device, electronic equipment and storage medium |
| WO2022099511A1 (en) * | 2020-11-11 | 2022-05-19 | 深圳元戎启行科技有限公司 | Method and apparatus for ground segmentation based on point cloud data, and computer device |
| CN115240149A (en) * | 2021-04-25 | 2022-10-25 | 株洲中车时代电气股份有限公司 | Three-dimensional point cloud detection and identification method and device, electronic equipment and storage medium |
-
2022
- 2022-11-16 CN CN202211436085.9A patent/CN115508807B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108052624A (en) * | 2017-12-15 | 2018-05-18 | 深圳市易成自动驾驶技术有限公司 | Processing Method of Point-clouds, device and computer readable storage medium |
| US10634793B1 (en) * | 2018-12-24 | 2020-04-28 | Automotive Research & Testing Center | Lidar detection device of detecting close-distance obstacle and method thereof |
| WO2021097618A1 (en) * | 2019-11-18 | 2021-05-27 | 深圳市大疆创新科技有限公司 | Point cloud segmentation method and system, and computer storage medium |
| WO2022099511A1 (en) * | 2020-11-11 | 2022-05-19 | 深圳元戎启行科技有限公司 | Method and apparatus for ground segmentation based on point cloud data, and computer device |
| CN112508912A (en) * | 2020-12-07 | 2021-03-16 | 中联重科股份有限公司 | Ground point cloud data filtering method and device and boom anti-collision method and system |
| CN115240149A (en) * | 2021-04-25 | 2022-10-25 | 株洲中车时代电气股份有限公司 | Three-dimensional point cloud detection and identification method and device, electronic equipment and storage medium |
| CN114219770A (en) * | 2021-11-26 | 2022-03-22 | 深圳市优必选科技股份有限公司 | Ground detection method, ground detection device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115508807B (en) | 2023-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7196412B2 (en) | Object recognition device and object recognition method | |
| CN114299417A (en) | Multi-target tracking method based on radar-vision fusion | |
| CN113160192A (en) | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background | |
| CN112949782A (en) | Target detection method, device, equipment and storage medium | |
| CN113688957A (en) | Target detection method, device, equipment and medium based on multi-model fusion | |
| CN112149476A (en) | Target detection method, device, equipment and storage medium | |
| CN112949519A (en) | Target detection method, device, equipment and storage medium | |
| CN114170487A (en) | Vision-based water surface oil stain detection method | |
| CN111860623A (en) | Method and system for counting tree number based on improved SSD neural network | |
| CN115508807B (en) | Point cloud data processing method and device, electronic equipment and storage medium | |
| CN113705672B (en) | Threshold selection method, system, device and storage medium for image target detection | |
| CN113610143B (en) | Method, device, equipment and storage medium for classifying point cloud noise points | |
| CN115984646A (en) | Distributed target detection method and device for remote sensing cross-satellite observation and satellite | |
| CN112686222B (en) | Method and system for detecting ship target by satellite-borne visible light detector | |
| CN115248443A (en) | Map construction method, system, equipment and storage medium based on laser radar | |
| CN114519400A (en) | Segmentation result evaluation method and device, electronic equipment and readable storage medium | |
| CN113963178A (en) | Method, device, equipment and medium for detecting infrared dim and small target under ground-air background | |
| JPH09152480A (en) | Automatic target recognition apparatus | |
| CN117593890B (en) | Detection method and device for road spilled objects, electronic equipment and storage medium | |
| CN117611592B (en) | Foreign matter detection method, device, electronic equipment and storage medium | |
| CN111476236B (en) | Self-adaptive FCM license plate positioning method and system | |
| CN115050016B (en) | License plate detection method and device, equipment terminal and readable storage medium | |
| CN117523428B (en) | Ground target detection method and device based on aircraft platform | |
| CN117237902B (en) | Robot character recognition system based on deep learning | |
| CN111368625B (en) | Pedestrian target detection method based on cascade optimization |
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 |