CN114638974A - Target object identification method, system, medium and electronic terminal - Google Patents
Target object identification method, system, medium and electronic terminal Download PDFInfo
- Publication number
- CN114638974A CN114638974A CN202210330855.5A CN202210330855A CN114638974A CN 114638974 A CN114638974 A CN 114638974A CN 202210330855 A CN202210330855 A CN 202210330855A CN 114638974 A CN114638974 A CN 114638974A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- dimensional
- cloud data
- data
- area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/521—Depth or shape recovery from laser ranging, e.g. using interferometry; from the projection of structured light
-
- 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 belongs to the field of warehouse logistics, and particularly relates to a target object identification method, a system, a medium and an electronic terminal, which are used for identifying a target object by acquiring three-dimensional point cloud data of an automobile parking area, then the three-dimensional point cloud data of the automobile parking area is identified and divided into three-dimensional point cloud data of the suspected area of the target object, thereby realizing the rough segmentation of the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil, acquiring the three-dimensional point cloud data or the two-dimensional outline data of the target object from the three-dimensional point cloud data of the suspected area of the target object by utilizing a random sampling consistency fitting algorithm, the target object is identified, so that subdivision and segmentation of the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil are achieved, and finally the steel coil and the saddle are identified by the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil, so that the saddle and the steel coil are identified simultaneously.
Description
Technical Field
The invention belongs to the field of warehouse logistics, and particularly relates to a target object identification method, a target object identification system, a target object identification medium and an electronic terminal.
Background
The steel production storage logistics transportation mainly comprises automobile, train and belt conveyor transportation, wherein the automobile transportation is the most flexible, and plays a vital role in the logistics of semi-finished product material transportation and outward logistics of finished product material production in the middle link of steel production.
The warehouse logistics in the steel industry enters an intelligent development stage, an intelligent warehouse system taking bridge crane unmanned driving as a core realizes unmanned operation of hoisting, loading and unloading warehouse materials, improves the operation efficiency and the essential safety of a warehouse area, and has a consensus in the industry.
The intelligent identification is carried out on the carrying automobile, and the accurate coordinate positions of a steel coil and a saddle on the carrying automobile are obtained, so that the requirement for completing unmanned loading and unloading of the automobile is met; however, in the past, based on a processing method of three-dimensional point cloud data, it was difficult to identify a saddle and a steel coil simultaneously in a carrier vehicle.
Disclosure of Invention
The invention provides a target object identification method, a target object identification system, a medium and an electronic terminal, and aims to solve the technical problem that steel coils and saddles on a delivery vehicle cannot be identified in the prior art.
A target object identification method, system, medium and electronic terminal includes the steps:
acquiring three-dimensional point cloud data of an automobile parking area;
carrying out feature recognition on the three-dimensional point cloud data of the automobile parking area to obtain the three-dimensional point cloud data of the suspected area of the target object;
determining three-dimensional point cloud data or two-dimensional contour data of a plurality of target objects from the three-dimensional point cloud data of the suspected target object areas;
and obtaining the characteristic information of all the target objects according to the three-dimensional point cloud data or the two-dimensional outline data of the plurality of target objects, and finishing the identification of all the target objects.
Optionally, the step of performing feature recognition on the three-dimensional point cloud data of the automobile parking area and acquiring the three-dimensional point cloud data of the suspected target area according to a feature recognition result includes:
establishing a conversion proportional relation between the height value and the gray level of the automobile parking area;
converting the height of the three-dimensional point cloud data of the automobile parking area into a gray value according to the conversion proportional relation to obtain a gray image and a two-dimensional-three-dimensional mapping relation;
and carrying out feature detection and segmentation on the gray level image to obtain a suspected area of the target object, and obtaining three-dimensional point cloud data of the suspected area of the target object according to the two-dimensional-three-dimensional mapping corresponding relation.
Optionally, the step of performing feature detection and segmentation on the grayscale image comprises:
acquiring an image containing a target object;
inputting the image containing the target object into an artificial neural network for training to generate a recognition model;
and carrying out feature detection and segmentation on the target object in the gray-scale image through the identification model to obtain a suspected area of the target object.
Optionally, the target suspected area includes a steel coil suspected area, and when the target suspected area is the steel coil suspected area, the steps of acquiring the three-dimensional point cloud data of the steel coil and acquiring the characteristic information of the steel coil from the three-dimensional point cloud data of the steel coil suspected area include:
detecting the three-dimensional point cloud data of the suspected area of the steel coil from bottom to top, and acquiring the three-dimensional point cloud data of a separated part when the three-dimensional point cloud data of the suspected area of the steel coil is detected to be separated;
acquiring three-dimensional point cloud data of the steel coil from the three-dimensional point cloud data of the separation part by adopting a random sampling consistent fitting algorithm;
and acquiring the characteristic information of the steel coil according to the three-dimensional point cloud data of the steel coil.
Optionally, the suspected target area includes a suspected saddle area, and when the suspected target area is the suspected saddle area, the steps of obtaining two-dimensional contour data of a saddle and obtaining feature information of the saddle according to the three-dimensional point cloud data of the suspected saddle area include:
projecting the three-dimensional point cloud data of the suspected saddle area along the side direction of the automobile to obtain two-dimensional outline data of the suspected saddle area;
detecting the two-dimensional contour data of the suspected saddle area from bottom to top, and acquiring the two-dimensional contour data of a separated part when the two-dimensional contour data of the suspected saddle area are separated;
acquiring two-dimensional profile data of the saddle baffle from the two-dimensional profile data of the separated part by adopting a random sampling consistent fitting algorithm;
splicing the two-dimensional profile data of the saddle baffle into the two-dimensional profile data of the saddle;
and acquiring characteristic information of the saddle according to the two-dimensional profile data of the saddle.
Optionally, the step of acquiring three-dimensional point cloud data of the automobile parking area comprises:
scanning the automobile parking area to obtain scanning data;
and establishing a three-dimensional coordinate system, performing three-dimensional reconstruction on the scanning data, rotating and translating the scanning data, and mapping the scanning data into the three-dimensional coordinate system to obtain three-dimensional point cloud data of the automobile parking area.
Optionally, the step of projecting the three-dimensionally reconstructed scanning data into the three-dimensional coordinate system to obtain three-dimensional point cloud data of the car parking area includes:
projecting the scanning data after three-dimensional reconstruction into the three-dimensional coordinate system to generate three-dimensional contour point cloud data;
removing discrete interference point data in the three-dimensional contour point cloud data through a statistical filtering algorithm;
and performing meshing filtration on the three-dimensional contour point cloud data after filtration, and converting the three-dimensional contour point cloud data into three-dimensional point cloud data of the automobile parking area.
The present invention also provides a target recognition system, comprising:
the acquisition module is used for acquiring three-dimensional point cloud data of an automobile parking area;
the first identification module is connected with the acquisition module and used for carrying out feature identification on the three-dimensional point cloud data of the automobile parking area and acquiring the three-dimensional point cloud data of the suspected area of the target object according to a feature identification result;
the separation module is connected with the identification module and used for determining three-dimensional point cloud data of the target object from the three-dimensional point cloud data of the suspected area of the target object;
and the second identification module is connected with the separation module and used for acquiring the characteristic information of all the target objects according to the three-dimensional point cloud data or the two-dimensional outline data of the plurality of target objects and completing the identification of all the target objects.
The invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method as defined in any one of the above.
The present invention also provides an electronic terminal, comprising: a processor and a memory;
the memory is adapted to store a computer program and the processor is adapted to execute the computer program stored by the memory to cause the terminal to perform the method as defined in any one of the above.
The invention provides a target object identification method, a system, a medium and an electronic terminal, which have the following beneficial effects: the saddle is divided into three-dimensional point cloud data of a suspected target area by acquiring the three-dimensional point cloud data of the automobile parking area and identifying the three-dimensional point cloud data of the automobile parking area, so that the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil are roughly divided, the three-dimensional point cloud data or the two-dimensional outline data of the target are acquired from the three-dimensional point cloud data of the suspected target area by using a random sampling consistency fitting algorithm, the target is identified, the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil are subdivided, and finally the steel coil and the saddle are identified by using the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil, so that the saddle and the steel coil are identified simultaneously.
Drawings
FIG. 1 is a schematic view of a car saddle and a steel coil identification method according to an embodiment of the present invention;
FIG. 2 is a schematic view of a vehicle parking area in an embodiment of the present invention;
FIG. 3 is a schematic diagram of a process for acquiring three-dimensional point cloud data of a parking area of a vehicle according to an embodiment of the invention;
FIG. 4 is a schematic diagram of a process for identifying and segmenting three-dimensional point cloud data of a parking area of a vehicle according to an embodiment of the invention;
FIG. 5 is a schematic diagram of a process for feature detection and segmentation of a grayscale image using an artificial neural network according to an embodiment of the present invention;
fig. 6 is a schematic diagram illustrating a process of separating three-dimensional point cloud data of a steel coil from three-dimensional point cloud data of a suspected area of the steel coil according to an embodiment of the invention;
FIG. 7 is a schematic representation of the process of separating two-dimensional saddle contour data from three-dimensional point cloud data of a suspected saddle area in an embodiment of the invention;
FIG. 8 is a schematic illustration of the result of processing three-dimensional point cloud data of a car parking area in an embodiment of the invention;
FIG. 9 is a schematic structural diagram of a car saddle and a steel coil identification system according to an embodiment of the present invention;
the reference numbers illustrate:
1 laser scanner
2 steel coil
3 saddle
4 computer
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
In the following description, numerous details are set forth to provide a more thorough explanation of embodiments of the present invention, however, it will be apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.
The inventor finds that when the bridge crane is used for loading and unloading the steel coil to the automobile, the automatic identification equipment can often receive the interference of saddles and automobile types, and the bridge crane can not obtain the accurate position of the steel coil, so that the steel coil is difficult to load and unload. Therefore, an efficient, accurate and high-robustness automobile saddle and a steel coil identification method are needed to solve the problems of accurate identification and measurement of various automobile types and complex saddle types in steel storage logistics transportation, improve the robustness of a storage sensing system and guarantee the storage logistics operation safety.
As shown in fig. 1, the method for identifying a target object provided in the present invention includes the steps of:
s1, acquiring three-dimensional point cloud data of an automobile parking area; the point cloud data (pointcloud data) refers to a set of vectors in a three-dimensional coordinate system, saddles, steel coils and other parts in an automobile parking area are recorded in the form of points, and each point comprises three-dimensional coordinates, and some points may comprise color information (RGB) or reflection Intensity information (Intensity);
as shown in fig. 2, in some embodiments, the parking area of the automobile comprises a parked automobile, and the automobile is respectively provided with a saddle 3 and a steel coil 2; a scanning device such as a laser scanner 1 is provided in a car stop area, and a computer 4 for processing various data is provided; the computer 4 analyzes the message sent by the laser scanner 1;
as shown in fig. 3, the step of acquiring the three-dimensional point cloud data of the automobile parking area includes:
s101, scanning an automobile parking area to obtain scanning data, wherein the scanning data is located in a visual coordinate system;
the laser scanner is positioned right above the automobile parking area, the laser scanner scans the automobile parking area, the generated data message is transmitted to the computer, and the computer analyzes the data message to generate scanning data;
s102, establishing a three-dimensional coordinate system, performing three-dimensional reconstruction on the scanning data, reconstructing the scanning data into three-dimensional data through the three-dimensional reconstruction, and mapping the scanning data subjected to the three-dimensional reconstruction into the three-dimensional coordinate system after rotating and translating to generate three-dimensional contour point cloud data of an automobile parking area. And based on the relative relation between the visual coordinate system and the three-dimensional coordinate system, rotating and translating the data under the visual coordinate system, and mapping the data into the three-dimensional coordinate system to obtain the data under the three-dimensional coordinate system. Since the laser scanner scans the car parking area from the top, the three-dimensional point cloud data of the car parking area can be used to represent the top contour of the car parking area.
S103, removing discrete interference point data in the three-dimensional contour point cloud data through a statistical filtering algorithm;
and S104, defining a grid density value, and carrying out meshing filtration on the filtered three-dimensional contour point cloud data according to the grid density value to form normalized three-dimensional point cloud data, namely the three-dimensional point cloud data of the automobile parking area.
S2, carrying out feature recognition on the three-dimensional point cloud data of the automobile parking area, and acquiring the three-dimensional point cloud data of the suspected steel coil area and the three-dimensional point cloud data of the suspected saddle area according to a feature recognition result; and (4) adopting characteristic identification to divide the three-dimensional point cloud data of the suspected area of the steel coil and the three-dimensional point cloud data of the suspected area of the saddle from the three-dimensional point cloud data of the parking area of the automobile.
As shown in fig. 4, in some embodiments, the identification and segmentation process includes:
s201, acquiring the relation between the height value and the ground plane of the automobile according to the three-dimensional point cloud data of the automobile parking area;
s202, setting a conversion proportional relation between the height and the gray level according to the relation between the height value of the automobile and the ground level;
s203, converting the height of the three-dimensional point cloud data of the automobile parking area into a gray value according to the conversion proportion relation to obtain a gray image and a two-dimensional-three-dimensional mapping relation;
in some embodiments, a vehicle height value and a ground plane can be obtained through three-dimensional point cloud data of a vehicle parking area, the height of the ground plane is set to be converted into a first gray value, the height corresponding to the vehicle height value is converted into a second gray value, and then the gray value is linearly given to an area between the ground plane and the top of the vehicle according to the conversion relation between the height and the gray value; obtaining a gray level image and a two-dimensional-three-dimensional mapping relation, wherein the two-dimensional-three-dimensional mapping relation is the corresponding relation of height and gray level value, and the gray level image can be restored into three-dimensional point cloud data according to the two-dimensional-three-dimensional mapping relation;
s204, carrying out feature detection and segmentation on the gray level image to obtain a steel coil suspected area and a saddle suspected area, and obtaining three-dimensional point cloud data of the steel coil suspected area and three-dimensional point cloud data of the saddle suspected area according to a two-dimensional-three-dimensional mapping corresponding relation;
detecting a suspected area of a steel coil and a suspected area of a saddle from the gray level image, but converting the suspected area of the steel coil and the suspected area of the saddle into three-dimensional point cloud data during further processing, so that the gray level characteristics in the suspected area of the steel coil and the suspected area of the saddle are reduced into height characteristics by utilizing a two-dimensional-three-dimensional mapping corresponding relation, and the three-dimensional point cloud data of the suspected area of the steel coil and the three-dimensional point cloud data of the suspected area of the saddle are obtained;
as shown in fig. 5, in some embodiments, the artificial neural network is used to perform feature detection and segmentation on the grayscale image, and the steps include:
s20401, acquiring steel coil graphic data and saddle graphic data to generate a training data set containing the steel coil graphic and the saddle graphic, and specifically generating the training data set aiming at the curved surface characteristic of the steel coil graphic and the outline characteristic of the saddle graphic;
s20402, establishing an artificial neural network, such as a YOLOv3 network, training the artificial neural network by using a training data set to generate a recognition model, and realizing automatic detection of a suspected steel coil area and a suspected saddle area by the artificial neural network through a large amount of training of steel coil graphs and saddle graphs;
s20403, carrying out feature detection and segmentation on the texture in the gray level image through an identification model, wherein the texture to be detected specifically comprises the curved surface feature of the steel coil (generally, the steel coil is cylindrical) and the contour feature of a saddle; the texture recognition of the gray level image is carried out through an artificial neural network, and the purpose is to roughly divide the area of a suspected steel coil and a saddle so as to be convenient for further processing.
S3, separating three-dimensional point cloud data of the steel coil from the three-dimensional point cloud data of the suspected steel coil area, and acquiring two-dimensional outline data of the saddle according to the three-dimensional point cloud data of the suspected saddle area; the purpose of the invention is to obtain the accurate positions of the steel coil and the saddle, so that the three-dimensional point cloud data of the suspected area of the steel coil and the suspected area of the saddle needs to be further accurately positioned to obtain the specific steel coil coordinates and saddle coordinates in the suspected area of the steel coil and the suspected area of the saddle.
As shown in fig. 6, in some embodiments, the step of separating the three-dimensional point cloud data of the steel coil from the three-dimensional point cloud data of the suspected area of the steel coil includes:
s301, scanning three-dimensional point cloud data of the suspected steel coil area from bottom to top from a first height, stopping scanning until the second height is reached, separating the three-dimensional point cloud data of the suspected steel coil area at the second height and above, and acquiring three-dimensional point cloud data of a separated part; the three-dimensional point cloud data of the automobile parking area acquired by the laser scanner is used for representing the outline of the top of the automobile parking area, so that the three-dimensional point cloud data of the bottom or the lower part of the automobile parking area are adhered together, the three-dimensional point cloud data of the steel coil is separated according to a bottom-up principle (namely, when the three-dimensional point cloud data of the suspected area of the steel coil is scanned from bottom to top, the three-dimensional point cloud data of the steel coil is necessarily separated), and the three-dimensional point cloud data of the separated part can be preliminarily judged to be the three-dimensional point cloud data of the steel coil.
S302, generating a steel coil model according to the user-defined information, and acquiring three-dimensional point cloud data of the steel coil from the three-dimensional point cloud data of the separated part by adopting a random sampling consistent fitting algorithm according to the steel coil model. The random sampling consistent fitting algorithm (RANDomSAmple Consenssus, RANSAC) estimates parameters of a mathematical model from a group of observed data containing outliers in an iterative mode, finds out three-dimensional point cloud data conforming to a steel coil model by using the random sampling consistent fitting algorithm to serve as the three-dimensional point cloud data of the steel coil, and can obtain characteristic information of the steel coil according to the three-dimensional point cloud data of the steel coil so as to obtain accurate coordinates of the steel coil.
As shown in fig. 7, in some embodiments, the step of separating the two-dimensional profile data of the saddle from the three-dimensional point cloud data of the suspected area of the saddle comprises:
s311, projecting the three-dimensional point cloud data of the suspected saddle area along the side direction of the automobile to obtain two-dimensional contour data of the suspected saddle area; three-dimensional point cloud data of a saddle suspected area are obtained through gray image identification, and then the three-dimensional point cloud data are projected, so that two-dimensional outline data representing the side outline of the saddle suspected area can be obtained;
s312, scanning the two-dimensional contour data from the third height to the fourth height from bottom to top, and separating the two-dimensional contour data at the fourth height to obtain two-dimensional contour data of a separated part; similarly, separating the two-dimensional profile data of the suspected saddle area according to a bottom-up principle; in particular, there is a possibility of placing a coil of steel on the saddle, but the edge portions (the ends of the coil of steel and the ends of the saddle) are generally not perfectly aligned and the coil of steel is shorter than the saddle, so that the top profile of the saddle can be obtained from the two-dimensional profile data of the suspected area of the saddle;
s313, generating a straight line model according to the user-defined information, and acquiring two-dimensional contour data of the saddle baffle from the two-dimensional contour data of the separated part by adopting a random sampling consistent fitting algorithm according to the straight line model; finding out two-dimensional contour data which accords with a linear model from the two-dimensional contour data of the suspected area of the saddle by utilizing a random sampling consistent fitting algorithm and is used for representing a saddle baffle (the saddle is composed of baffles, and the side projection of the baffles is a straight line)
S314, splicing the two-dimensional profile data of the saddle baffles into two-dimensional profile data of the saddle, finding out the two-dimensional profile data of all the saddle baffles, splicing the two-dimensional profile data to obtain the two-dimensional profile data of the saddle, and obtaining the characteristic information of the saddle according to the two-dimensional profile data of the saddle so as to obtain the accurate coordinates of the saddle.
And S4, identifying the target object according to the three-dimensional point cloud data or the two-dimensional outline data of the target object, specifically, acquiring the characteristic information of the steel coil according to the three-dimensional point cloud data of the steel coil, and acquiring the characteristic information of the saddle according to the two-dimensional outline data of the saddle. The steel coil and the saddle are more accurately identified through the characteristic information of the steel coil and the characteristic information of the saddle, wherein the characteristic information of the steel coil comprises the outer diameter of the steel coil, the width dimension of the steel coil, the normal vector of the axis of the steel coil, the coordinate of the central point of the axis and the like; the characteristic information of the saddle comprises the included angle of saddle baffles, the distance between saddle baffles, the size of the saddle baffles and the like.
And finally, the characteristic information of the steel coil and the saddle is fed back to the warehousing system, so that the steel coil can be precisely hoisted.
The invention provides a target object identification method, which comprises the steps of acquiring three-dimensional point cloud data of an automobile parking area, identifying the three-dimensional point cloud data of the automobile parking area, and dividing the three-dimensional point cloud data into three-dimensional point cloud data of a suspected target object area; specifically, three-dimensional point cloud data of an automobile parking area are transformed through a height-to-gray scale conversion proportional relation, features in the height direction are converted into gray scale feature descriptions on gray scale images, then steel coil curved surface features and saddle outline feature detection are carried out on the gray scale images through an artificial neural network, segmentation of a steel coil suspected area and the saddle suspected area is achieved, three-dimensional point cloud data or two-dimensional outline data of a target object are obtained from the three-dimensional point cloud data of the target object suspected area through a random sampling consistency fitting algorithm, the target object is identified, accordingly, segmentation of the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil is achieved, finally the steel coil and the saddle are identified through the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil, and accordingly, and the saddle and the steel coil are identified at the same time; and successively segmenting, fitting and screening the suspected area data of the steel coil and the saddle to obtain the characteristic information of the steel coil and the saddle. The detection precision and efficiency are improved, the system robustness is improved, and the safety of warehouse logistics operation is guaranteed.
As shown in fig. 9, the present invention also provides a target recognition system, including:
the acquisition module is used for acquiring three-dimensional point cloud data of an automobile parking area;
the acquisition module is used for acquiring three-dimensional point cloud data of an automobile parking area;
the first identification module is connected with the acquisition module and used for carrying out feature identification on the three-dimensional point cloud data of the automobile parking area and acquiring the three-dimensional point cloud data of the suspected area of the target object according to a feature identification result;
the separation module is connected with the identification module and used for determining three-dimensional point cloud data of the target object from the three-dimensional point cloud data of the suspected area of the target object;
and the second identification module is connected with the separation module and used for acquiring the characteristic information of all the target objects according to the three-dimensional point cloud data or the two-dimensional outline data of the plurality of target objects and completing the identification of all the target objects.
The invention provides a target object identification system, which is characterized in that three-dimensional point cloud data of an automobile parking area are obtained, then the three-dimensional point cloud data of the automobile parking area are identified and divided into three-dimensional point cloud data of a suspected target object area; specifically, three-dimensional point cloud data of an automobile parking area are transformed through a height-to-gray scale conversion proportional relation, features in the height direction are converted into gray scale feature descriptions on gray scale images, then steel coil curved surface features and saddle outline feature detection are carried out on the gray scale images through an artificial neural network, segmentation of a steel coil suspected area and the saddle suspected area is achieved, three-dimensional point cloud data or two-dimensional outline data of a target object are obtained from the three-dimensional point cloud data of the target object suspected area through a random sampling consistency fitting algorithm, the target object is identified, accordingly, segmentation of the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil is achieved, finally the steel coil and the saddle are identified through the two-dimensional outline data of the saddle and the three-dimensional point cloud data of the steel coil, and accordingly, and the saddle and the steel coil are identified at the same time; and successively segmenting, fitting and screening the suspected area data of the steel coil and the saddle to obtain the characteristic information of the steel coil and the saddle. The detection precision and efficiency are improved, the system robustness is improved, and the safety of warehouse logistics operation is guaranteed.
The present embodiment also provides a computer-readable storage medium on which a computer program is stored, which when executed by a processor implements any of the methods in the present embodiments.
The present embodiment further provides an electronic terminal, including: a processor and a memory;
the memory is used for storing computer programs, and the processor is used for executing the computer programs stored in the memory so as to enable the terminal to execute the method in the embodiment.
The computer-readable storage medium in the present embodiment can be understood by those skilled in the art as follows: all or part of the steps for implementing the above method embodiments may be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The electronic terminal provided by the embodiment comprises a processor, a memory, a transceiver and a communication interface, wherein the memory and the communication interface are connected with the processor and the transceiver and are used for completing mutual communication, the memory is used for storing a computer program, the communication interface is used for carrying out communication, and the processor and the transceiver are used for operating the computer program so that the electronic terminal can execute the steps of the method.
In this embodiment, the Memory may include a Random Access Memory (RAM), and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the Integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
In the embodiments described above, although the present invention has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those skilled in the art in light of the foregoing description. The embodiments of the invention are intended to embrace all such alternatives, modifications and variances that fall within the broad scope of the appended claims.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (10)
1. A method for identifying an object, comprising the steps of:
acquiring three-dimensional point cloud data of an automobile parking area;
carrying out feature recognition on the three-dimensional point cloud data of the automobile parking area to obtain the three-dimensional point cloud data of the suspected area of the target object;
determining three-dimensional point cloud data or two-dimensional contour data of a plurality of target objects from the three-dimensional point cloud data of the suspected target object areas;
and obtaining the characteristic information of all the target objects according to the three-dimensional point cloud data or the two-dimensional outline data of the plurality of target objects, and finishing the identification of all the target objects.
2. The method for identifying the target object according to claim 1, wherein the step of performing feature identification on the three-dimensional point cloud data of the automobile parking area and acquiring the three-dimensional point cloud data of the suspected target object area according to a feature identification result comprises:
establishing a conversion proportional relation between the height value and the gray level of the automobile parking area;
converting the height of the three-dimensional point cloud data of the automobile parking area into a gray value according to the conversion proportional relation to obtain a gray image and a two-dimensional-three-dimensional mapping relation;
and carrying out feature detection and segmentation on the gray level image to obtain a suspected area of the target object, and obtaining three-dimensional point cloud data of the suspected area of the target object according to the two-dimensional-three-dimensional mapping corresponding relation.
3. The method of claim 2, wherein the step of performing feature detection and segmentation on the gray scale image comprises:
acquiring an image containing a target object;
inputting the image containing the target object into an artificial neural network for training to generate a recognition model;
and carrying out feature detection and segmentation on the target object in the gray-scale image through the identification model to obtain a suspected area of the target object.
4. The method according to claim 1, wherein the suspected target area includes a suspected steel coil area, and the steps of obtaining three-dimensional point cloud data of a steel coil and obtaining characteristic information of the steel coil from the three-dimensional point cloud data of the suspected steel coil area when the suspected target area is the suspected steel coil area include:
detecting the three-dimensional point cloud data of the suspected steel coil area from bottom to top, and acquiring the three-dimensional point cloud data of a separation part when detecting the separation of the three-dimensional point cloud data of the suspected steel coil area;
acquiring three-dimensional point cloud data of the steel coil from the three-dimensional point cloud data of the separation part by adopting a random sampling consistent fitting algorithm;
and acquiring the characteristic information of the steel coil according to the three-dimensional point cloud data of the steel coil.
5. The method according to claim 1, wherein the suspected target area comprises a saddle suspected area, and when the suspected target area is the saddle suspected area, the steps of obtaining two-dimensional profile data of a saddle and obtaining saddle feature information from three-dimensional point cloud data of the saddle suspected area comprise:
projecting the three-dimensional point cloud data of the suspected saddle area along the side direction of the automobile to obtain two-dimensional contour data of the suspected saddle area;
detecting the two-dimensional contour data of the suspected saddle area from bottom to top, and acquiring the two-dimensional contour data of a separated part when the two-dimensional contour data of the suspected saddle area are separated;
acquiring two-dimensional profile data of the saddle baffle from the two-dimensional profile data of the separated part by adopting a random sampling consistent fitting algorithm;
splicing the two-dimensional profile data of the saddle baffle into the two-dimensional profile data of the saddle;
and acquiring characteristic information of the saddle according to the two-dimensional profile data of the saddle.
6. The method for identifying the target object according to claim 1, wherein the step of obtaining the three-dimensional point cloud data of the parking area of the automobile comprises:
scanning the automobile parking area to obtain scanning data;
and establishing a three-dimensional coordinate system, performing three-dimensional reconstruction on the scanning data, rotating and translating the scanning data, and mapping the scanning data into the three-dimensional coordinate system to obtain three-dimensional point cloud data of the automobile parking area.
7. The method as claimed in claim 6, wherein the step of projecting the three-dimensionally reconstructed scan data into the three-dimensional coordinate system to obtain the three-dimensional point cloud data of the car parking area comprises:
projecting the scanning data after three-dimensional reconstruction into the three-dimensional coordinate system to generate three-dimensional contour point cloud data;
removing discrete interference point data in the three-dimensional contour point cloud data through a statistical filtering algorithm;
and performing meshing filtration on the three-dimensional contour point cloud data after filtration, and converting the three-dimensional contour point cloud data into three-dimensional point cloud data of the automobile parking area.
8. An object recognition system, comprising:
the acquisition module is used for acquiring three-dimensional point cloud data of an automobile parking area;
the first identification module is connected with the acquisition module and used for carrying out feature identification on the three-dimensional point cloud data of the automobile parking area and acquiring the three-dimensional point cloud data of the suspected area of the target object according to a feature identification result;
the separation module is connected with the identification module and used for determining three-dimensional point cloud data of the target object from the three-dimensional point cloud data of the suspected area of the target object;
and the second identification module is connected with the separation module and used for acquiring the characteristic information of all the target objects according to the three-dimensional point cloud data or the two-dimensional outline data of the plurality of target objects and completing the identification of all the target objects.
9. A computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program, when executed by a processor, implements the method of any one of claims 1 to 7.
10. An electronic terminal, comprising: a processor and a memory;
the memory is for storing a computer program and the processor is for executing the computer program stored by the memory to cause the terminal to perform the method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210330855.5A CN114638974A (en) | 2022-03-29 | 2022-03-29 | Target object identification method, system, medium and electronic terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210330855.5A CN114638974A (en) | 2022-03-29 | 2022-03-29 | Target object identification method, system, medium and electronic terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114638974A true CN114638974A (en) | 2022-06-17 |
Family
ID=81951572
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210330855.5A Pending CN114638974A (en) | 2022-03-29 | 2022-03-29 | Target object identification method, system, medium and electronic terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114638974A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797444A (en) * | 2023-02-06 | 2023-03-14 | 中国科学院自动化研究所 | Pineapple eye positioning method and device and electronic equipment |
| CN117784169A (en) * | 2024-02-27 | 2024-03-29 | 唐山港集团股份有限公司 | 3D point cloud-based steel coil contour measurement method, equipment and medium |
-
2022
- 2022-03-29 CN CN202210330855.5A patent/CN114638974A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797444A (en) * | 2023-02-06 | 2023-03-14 | 中国科学院自动化研究所 | Pineapple eye positioning method and device and electronic equipment |
| CN117784169A (en) * | 2024-02-27 | 2024-03-29 | 唐山港集团股份有限公司 | 3D point cloud-based steel coil contour measurement method, equipment and medium |
| CN117784169B (en) * | 2024-02-27 | 2024-05-07 | 唐山港集团股份有限公司 | 3D point cloud-based steel coil contour measurement method, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114638974A (en) | Target object identification method, system, medium and electronic terminal | |
| CN110751620B (en) | Method for estimating volume and weight, electronic device, and computer-readable storage medium | |
| CN111968181B (en) | Real-time parcel position detection method and system based on image processing | |
| CN111767780B (en) | AI and vision combined intelligent integrated card positioning method and system | |
| CN107609510B (en) | Positioning method and device for lower set of quayside container crane | |
| CN109523527A (en) | The detection method in dirty region, device and electronic equipment in image | |
| CN107527339B (en) | Magnetic resonance scanning method, device and system | |
| CN109993758A (en) | Dividing method, segmenting device, computer equipment and storage medium | |
| Yogeswaran et al. | 3d surface analysis for automated detection of deformations on automotive body panels | |
| CN111582000A (en) | Bar code positioning method and device and related equipment | |
| CN115760820A (en) | Plastic part defect image identification method and application | |
| CN115937203A (en) | Visual detection method, device, equipment and medium based on template matching | |
| JPH07220090A (en) | Object recognition method | |
| CN114742789B (en) | General part picking method and system based on surface structured light and electronic equipment | |
| CN113469971A (en) | Image matching method, image matching detection device and storage medium | |
| CN115546300A (en) | Method and device for identifying pose of tray placed tightly, computer equipment and medium | |
| CN113435219B (en) | Anti-counterfeiting detection method and device, electronic equipment and storage medium | |
| EP3846127A1 (en) | Dark parcel dimensioning | |
| CN111582778B (en) | Method, device, equipment and storage medium for measuring accumulation of cargos in operation site | |
| CN113610772A (en) | Method, system and device for detecting bottle bottom code spraying defects of pop cans and storage medium | |
| CN116843742B (en) | Calculation method and system for stacking volume after point cloud registration for black coal loading vehicle | |
| CN115631329B (en) | Loading control method, system and storage medium for open carriage | |
| CN110751732A (en) | Method for converting 2D image into 3D image | |
| CN111415378A (en) | Image registration method for automobile glass detection and automobile glass detection method | |
| CN113978987B (en) | Pallet object packaging and picking method, device, equipment and medium |
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 | ||
| CB02 | Change of applicant information |
Address after: 401329 No. 5-6, building 2, No. 66, Nongke Avenue, Baishiyi Town, Jiulongpo District, Chongqing Applicant after: MCC CCID information technology (Chongqing) Co.,Ltd. Address before: 401329 No. 5-6, building 2, No. 66, Nongke Avenue, Baishiyi Town, Jiulongpo District, Chongqing Applicant before: CISDI CHONGQING INFORMATION TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information |