CN117036424A - Continuous frame point cloud labeling method and system based on registration - Google Patents

Continuous frame point cloud labeling method and system based on registration Download PDF

Info

Publication number
CN117036424A
CN117036424A CN202310836010.8A CN202310836010A CN117036424A CN 117036424 A CN117036424 A CN 117036424A CN 202310836010 A CN202310836010 A CN 202310836010A CN 117036424 A CN117036424 A CN 117036424A
Authority
CN
China
Prior art keywords
frame
target
point cloud
point
target object
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
Application number
CN202310836010.8A
Other languages
Chinese (zh)
Inventor
李森林
李诒雯
郝江波
谢赤天
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan Kotei Informatics Co Ltd
Original Assignee
Wuhan Kotei Informatics Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan Kotei Informatics Co Ltd filed Critical Wuhan Kotei Informatics Co Ltd
Priority to CN202310836010.8A priority Critical patent/CN117036424A/en
Publication of CN117036424A publication Critical patent/CN117036424A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds

Abstract

The invention relates to a continuous frame point cloud labeling method and a continuous frame point cloud labeling system based on registration, wherein the method comprises the following steps: acquiring a target object labeling frame of a previous frame of a target frame and a point cloud point set in a target object bounding frame; obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame; registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after successful registration; acquiring a target object labeling frame of the target frame based on the position matching relation and the target object labeling frame of the previous frame; by the point cloud registration concept, the accurate position of the target object labeling frame in the next frame is calculated, the manual adjustment time is greatly reduced, and the point cloud continuous frame labeling efficiency is improved.

Description

Continuous frame point cloud labeling method and system based on registration
Technical Field
The invention relates to the field of point cloud labeling, in particular to a continuous frame point cloud labeling method and system based on registration.
Background
With the continuous development of laser radar technology, laser radar has been widely applied to the fields of topographic mapping, ocean exploration, fishery resource monitoring, automatic driving and the like.
Labeling the target object in the laser point cloud can provide a data set for downstream tasks. The continuous frame labeling is an important field in point cloud labeling, the existing continuous frame point cloud labeling algorithm inherits the position of a target object labeling frame of a current frame into a second frame, and then manual adjustment is carried out by labeling personnel, so that the continuous frame labeling method is low in efficiency.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides a continuous frame point cloud labeling method and a continuous frame point cloud labeling system based on registration, and the accurate position of a target object labeling frame in a subsequent frame is calculated through the point cloud registration idea, so that the time of manual adjustment is greatly reduced, and the point cloud continuous frame labeling efficiency is improved. According to a first aspect of the present invention, there is provided a registration-based continuous frame point cloud labeling method, comprising: step 1, obtaining a target object labeling frame and a point cloud point set in a target object bounding frame of a previous frame of a target frame;
step 2, obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame;
step 3, registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful;
and step 4, acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
On the basis of the technical scheme, the invention can also make the following improvements.
Optionally, in step 1, when the previous frame of the target frame is the first frame, the standard frame of the previous frame is obtained by means of manual labeling.
Optionally, the obtaining the point cloud set in the target bounding box in step 1 includes:
acquiring the length l1, the width w1, the height h1 and the center point coordinates (x 1, y1, z 1) of each object bounding box;
obtaining eight vertex coordinates of a target object bounding box;
and (3) comparing the coordinates of the point cloud points of the traversing point cloud file with the vertex coordinates of each object bounding box, and determining a point cloud point set A of the object of the previous frame.
Optionally, the process of acquiring the point cloud point set in the object bounding box of the target frame in the step 2 includes:
and determining the same coordinates (x 2, y2, z 2) in the target frame as target datum points based on the central point coordinates (x 1, y1, z 1) of the target bounding box of the previous frame of the target frame, taking the datum points as the centers in the ranges of the length l1+4, the width w1+2 and the height h1, and obtaining a point cloud point set B in the target bounding box of the target frame.
Optionally, in step 3, an ICP point cloud registration algorithm is used to register the point clouds of the target frame and the previous frame, and if the registration accuracy is higher than a set value, the successful registration is determined.
Optionally, the matching relationship between the target frame and the point cloud of the previous frame obtained in the step 3 includes: a rotation matrix R and a translation matrix T representing the positional relationship of two point cloud point sets.
Optionally, step 4 includes:
and mapping the target object labeling frame of the previous frame to the target frame based on the rotation matrix R and the translation matrix T to obtain the target object labeling frame of the target frame.
According to a second aspect of the present invention, there is provided a registration-based continuous frame point cloud labeling system, comprising: the system comprises a previous frame point cloud point set acquisition module, a target frame point cloud point set acquisition module, a registration module and a target frame annotation frame acquisition module;
the previous frame point cloud point set acquisition module is used for acquiring a target object labeling frame of a previous frame of a target frame and a point cloud point set in the target object bounding frame;
the target frame point cloud point set acquisition module is used for acquiring a target object bounding box of a target frame based on a target object bounding box of a previous frame and acquiring a point cloud point set in the target object bounding box of the target frame;
the registration module is used for registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful;
the target frame annotation frame acquisition module is used for acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
According to a third aspect of the present invention, there is provided an electronic device comprising a memory, a processor for implementing the steps of a registration-based continuous frame point cloud labeling method when executing a computer management class program stored in the memory.
According to a fourth aspect of the present invention, there is provided a computer readable storage medium having stored thereon a computer management class program which when executed by a processor performs the steps of a registration-based continuous frame point cloud labeling method.
According to the continuous frame point cloud labeling method, the system, the electronic equipment and the storage medium based on registration, the point cloud registration idea is used, the rotation translation relation of the continuous frame point cloud target is obtained based on the characteristics of fixed form, small displacement and the like of the continuous frame point cloud target, the surrounding frame position of the next frame target is calculated based on the current frame, the accuracy and the labeling efficiency of continuous frame point cloud labeling are improved, and the manual operation complexity is reduced. The method can solve the problems of high manual participation and low precision in the traditional point cloud continuous frame labeling, control the registration precision threshold value by using a point cloud registration algorithm, ensure the prediction precision of the point cloud labeling frame, improve the automation degree of the point cloud continuous frame labeling and improve the labeling efficiency. .
Drawings
FIG. 1 is a flowchart of a registration-based continuous frame point cloud labeling method provided by the invention;
FIG. 2 is a block diagram of a registration-based continuous frame point cloud labeling system;
fig. 3 is a schematic hardware structure of one possible electronic device according to the present invention;
fig. 4 is a schematic hardware structure of a possible computer readable storage medium according to the present invention.
Detailed Description
The principles and features of the present invention are described below with reference to the drawings, the examples are illustrated for the purpose of illustrating the invention and are not to be construed as limiting the scope of the invention.
Fig. 1 is a flowchart of a registration-based continuous frame point cloud labeling method, as shown in fig. 1, where the point cloud labeling method includes:
step 1, obtaining a target object labeling frame of a frame before a target frame and a point cloud point set in the target object bounding frame.
And 2, obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame.
And 3, registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful.
And step 4, acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
According to the continuous frame point cloud labeling method based on registration, through the point cloud registration idea, the accurate position of the target object labeling frame in the next frame is calculated, the time of manual adjustment is greatly reduced, and the point cloud continuous frame labeling efficiency is improved.
Example 1
The embodiment 1 provided by the present invention is an embodiment of a registration-based continuous frame point cloud labeling method provided by the present invention, and as can be known from fig. 1, the embodiment of the point cloud labeling method includes:
step 1, obtaining a target object labeling frame of a frame before a target frame and a point cloud point set in the target object bounding frame.
In a possible embodiment, in step 1, when the previous frame of the target frame is the first frame, the standard frame of the previous frame is obtained by means of manual labeling.
In a possible embodiment, the acquiring the point cloud set in the object bounding box in step 1 includes:
the length l1, width w1, height h1 and center point coordinates (x 1, y1, z 1) of each object bounding box are acquired.
Eight vertex coordinates of the object bounding box are acquired.
And (3) comparing the coordinates of the point cloud points of the traversing point cloud file with the vertex coordinates of each object bounding box, and determining a point cloud point set A of the object of the previous frame.
And 2, obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame.
In a possible embodiment, the process of obtaining the set of points cloud points in the object bounding box of the object frame in step 2 includes:
and determining the same coordinates (x 2, y2, z 2) in the target frame as target datum points based on the central point coordinates (x 1, y1, z 1) of the target bounding box of the previous frame of the target frame, taking the datum points as the centers in the ranges of the length l1+4, the width w1+2 and the height h1, and obtaining a point cloud point set B in the target bounding box of the target frame.
And 3, registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful.
In a possible embodiment, in step 3, an ICP (Iterative Closest Point) point cloud registration algorithm is used to register the point clouds of the target frame and the frame before the target frame, and if the registration accuracy is higher than a set value, it is determined that the registration is successful.
The set point may be 95%.
In a possible embodiment, the matching relationship between the location of the point cloud of the target frame and the point cloud of the previous frame obtained in the step 3 includes: a rotation matrix R and a translation matrix T representing the positional relationship of two point cloud point sets.
And step 4, acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
In one possible embodiment, step 4 includes:
and mapping the target object labeling frame of the previous frame to the target frame based on the rotation matrix R and the translation matrix T to obtain the target object labeling frame of the target frame, namely finishing point cloud labeling of the continuous frame.
Example 2
Embodiment 2 provided by the present invention is an embodiment of a registration-based continuous frame point cloud labeling system provided by the present invention, and fig. 2 is a block diagram of the registration-based continuous frame point cloud labeling system provided by the embodiment of the present invention, as can be known from fig. 2, the embodiment of the point cloud labeling system includes:
the system comprises a previous frame point cloud point set acquisition module, a target frame point cloud point set acquisition module, a registration module and a target frame annotation frame acquisition module.
And the previous frame point cloud point set acquisition module is used for acquiring the target object labeling frame of the previous frame of the target frame and the point cloud point set in the target object bounding frame.
And the target frame point cloud point set acquisition module is used for acquiring a target object bounding box of the target frame based on the target object bounding box of the previous frame and acquiring a point cloud point set in the target object bounding box of the target frame.
The registration module is used for registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful.
The target frame annotation frame acquisition module is used for acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
It can be understood that the continuous frame point cloud labeling system based on registration provided by the present invention corresponds to the continuous frame point cloud labeling method based on registration provided in the foregoing embodiments, and relevant technical features of the continuous frame point cloud labeling system based on registration may refer to relevant technical features of the continuous frame point cloud labeling method based on registration, which are not described herein again.
Referring to fig. 3, fig. 3 is a schematic diagram of an embodiment of an electronic device according to an embodiment of the invention. As shown in fig. 3, an embodiment of the present invention provides an electronic device, including a memory 1310, a processor 1320, and a computer program 1311 stored in the memory 1310 and executable on the processor 1320, wherein the processor 1320 executes the computer program 1311 to implement the following steps: acquiring a target object labeling frame of a previous frame of a target frame and a point cloud point set in a target object bounding frame; obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame; registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after successful registration; and acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
Referring to fig. 4, fig. 4 is a schematic diagram of an embodiment of a computer readable storage medium according to the present invention. As shown in fig. 4, the present embodiment provides a computer-readable storage medium 1400 having stored thereon a computer program 1411, which computer program 1411, when executed by a processor, performs the steps of: acquiring a target object labeling frame of a previous frame of a target frame and a point cloud point set in a target object bounding frame; obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame; registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after successful registration; and acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
According to the continuous frame point cloud labeling method, the system, the electronic equipment and the storage medium based on registration, the point cloud registration idea is used, the rotation translation relation of the continuous frame point cloud target is obtained based on the characteristics of fixed form, small displacement and the like of the continuous frame point cloud target, the surrounding frame position of the next frame target is calculated based on the current frame, the accuracy and the labeling efficiency of continuous frame point cloud labeling are improved, and the manual operation complexity is reduced. The method can solve the problems of high manual participation and low precision in the traditional point cloud continuous frame labeling, control the registration precision threshold value by using a point cloud registration algorithm, ensure the prediction precision of the point cloud labeling frame, improve the automation degree of the point cloud continuous frame labeling and improve the labeling efficiency.
In the foregoing embodiments, the descriptions of the embodiments are focused on, and for those portions of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims (10)

1. The continuous frame point cloud labeling method based on registration is characterized by comprising the following steps of:
step 1, obtaining a target object labeling frame and a point cloud point set in a target object bounding frame of a previous frame of a target frame;
step 2, obtaining a target object bounding box of the target frame based on the target object bounding box of the previous frame, and obtaining a point cloud point set in the target object bounding box of the target frame;
step 3, registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful;
and step 4, acquiring the target object labeling frame of the target frame based on the position matching relation and the target object labeling frame of the previous frame.
2. The point cloud labeling method according to claim 1, wherein in the step 1, when a previous frame of the target frame is a first frame, a standard frame of the previous frame is obtained by means of manual labeling.
3. The method for labeling point clouds according to claim 1, wherein the step 1 of obtaining the point cloud set in the object bounding box comprises:
acquiring the length l1, the width w1, the height h1 and the center point coordinates (x 1, y1, z 1) of each object bounding box;
obtaining eight vertex coordinates of a target object bounding box;
and comparing the coordinates of the point cloud points of the traversing point cloud file with the vertex coordinates of the bounding boxes of each target object, and determining a point cloud point set A of the target object of the previous frame.
4. The method of point cloud labeling as set forth in claim 3, wherein the process of obtaining the point cloud point set in the object bounding box of the object frame in step 2 includes:
and determining the same coordinates (x 2, y2, z 2) in the target frame as target datum points based on the central point coordinates (x 1, y1, z 1) of the target bounding box of the previous frame of the target frame, taking the datum points as the center, and taking points in the ranges of the length l1+4, the width w1+2 and the height h1 to obtain a target bounding box inner point cloud point set B of the target frame.
5. The point cloud labeling method according to claim 1, wherein in the step 3, an ICP point cloud registration algorithm is used to register the point clouds of the target frame and the previous frame, and if the registration accuracy is higher than a set value, it is determined that the registration is successful.
6. The point cloud labeling method according to claim 1, wherein the position matching relationship of the point cloud of the target frame and the point cloud of the previous frame obtained in the step 3 comprises: a rotation matrix R and a translation matrix T representing the positional relationship of two point cloud point sets.
7. The point cloud labeling method of claim 1, wherein the step 4 comprises:
and mapping the target object labeling frame of the previous frame to the target frame based on the rotation matrix R and the translation matrix T to obtain the target object labeling frame of the target frame.
8. A registration-based continuous frame point cloud labeling system, the point cloud labeling system comprising: the system comprises a previous frame point cloud point set acquisition module, a target frame point cloud point set acquisition module, a registration module and a target frame annotation frame acquisition module;
the previous frame point cloud point set acquisition module is used for acquiring a target object labeling frame of a previous frame of a target frame and a point cloud point set in a target object bounding frame;
the target frame point cloud point set acquisition module is used for acquiring a target object bounding box of the target frame based on the target object bounding box of the previous frame and acquiring a point cloud point set in the target object bounding box of the target frame;
the registration module is used for registering the point clouds of the target frame and the previous frame, and acquiring the position matching relation of the point clouds of the target frame and the previous frame after the registration is successful;
the target frame annotation frame acquisition module is used for acquiring the target object annotation frame of the target frame based on the position matching relation and the target object annotation frame of the previous frame.
9. An electronic device comprising a memory, a processor for implementing the steps of the registration-based continuous frame point cloud labeling method of any of claims 1-7 when executing a computer management class program stored in the memory.
10. A computer readable storage medium, having stored thereon a computer management class program which when executed by a processor implements the steps of the registration-based continuous frame point cloud labeling method of any of claims 1-7.
CN202310836010.8A 2023-07-08 2023-07-08 Continuous frame point cloud labeling method and system based on registration Pending CN117036424A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310836010.8A CN117036424A (en) 2023-07-08 2023-07-08 Continuous frame point cloud labeling method and system based on registration

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310836010.8A CN117036424A (en) 2023-07-08 2023-07-08 Continuous frame point cloud labeling method and system based on registration

Publications (1)

Publication Number Publication Date
CN117036424A true CN117036424A (en) 2023-11-10

Family

ID=88638014

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310836010.8A Pending CN117036424A (en) 2023-07-08 2023-07-08 Continuous frame point cloud labeling method and system based on registration

Country Status (1)

Country Link
CN (1) CN117036424A (en)

Similar Documents

Publication Publication Date Title
CN111311632B (en) Object pose tracking method, device and equipment
CN110458772B (en) Point cloud filtering method and device based on image processing and storage medium
CN110599489A (en) Target space positioning method
CN109558471B (en) Updating method, device, storage medium and system of grid map
CN111105452B (en) Binocular vision-based high-low resolution fusion stereo matching method
CN109543634B (en) Data processing method and device in positioning process, electronic equipment and storage medium
CN111210475A (en) Map updating method and device
CN115810133B (en) Welding control method based on image processing and point cloud processing and related equipment
CN115329111B (en) Image feature library construction method and system based on point cloud and image matching
CN111340834B (en) Lining plate assembly system and method based on laser radar and binocular camera data fusion
CN115164868A (en) Robot positioning method, device, robot and storage medium
CN115326051A (en) Positioning method and device based on dynamic scene, robot and medium
CN110717141A (en) Lane line optimization method and device and storage medium
CN116543134B (en) Method, device, computer equipment and medium for constructing digital twin model
CN110363847B (en) Map model construction method and device based on point cloud data
CN117036424A (en) Continuous frame point cloud labeling method and system based on registration
CN116503760A (en) Unmanned aerial vehicle cruising detection method based on self-adaptive edge feature semantic segmentation
CN113762397B (en) Method, equipment, medium and product for training detection model and updating high-precision map
CN114331919B (en) Depth recovery method, electronic device, and storage medium
CN113808142A (en) Ground identifier identification method and device and electronic equipment
CN111260723A (en) Barycenter positioning method of bar and terminal equipment
CN114485608B (en) Local point cloud rapid registration method for high-precision map making
CN114022501A (en) Arrow corner automatic detection method and system, electronic equipment and storage medium
CN117994329A (en) Target positioning method and system based on stereoscopic vision point cloud
CN114445568A (en) Detection and extraction method and system for straight turning composite arrow

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