CN114187579A - Target detection method, apparatus and computer-readable storage medium for automatic driving - Google Patents

Target detection method, apparatus and computer-readable storage medium for automatic driving Download PDF

Info

Publication number
CN114187579A
CN114187579A CN202111525882.XA CN202111525882A CN114187579A CN 114187579 A CN114187579 A CN 114187579A CN 202111525882 A CN202111525882 A CN 202111525882A CN 114187579 A CN114187579 A CN 114187579A
Authority
CN
China
Prior art keywords
target object
dimensional
image
coordinate system
point cloud
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
CN202111525882.XA
Other languages
Chinese (zh)
Inventor
单国航
贾双成
朱磊
李成军
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.)
Zhidao Network Technology Beijing Co Ltd
Original Assignee
Zhidao Network Technology Beijing 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 Zhidao Network Technology Beijing Co Ltd filed Critical Zhidao Network Technology Beijing Co Ltd
Priority to CN202111525882.XA priority Critical patent/CN114187579A/en
Publication of CN114187579A publication Critical patent/CN114187579A/en
Pending legal-status Critical Current

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
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/86—Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Electromagnetism (AREA)
  • Traffic Control Systems (AREA)

Abstract

The application relates to an object detection method, device and computer readable storage medium for automatic driving. The method comprises the following steps: acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, wherein the driving environment image at least comprises an image of the surrounding environment when the intelligent vehicle drives; under the condition that a target object in any one of at least one driving environment image or three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object to obtain a projected region of interest; and detecting a target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interesting area and has a range larger than the projection interesting area. The scheme provided by the application can be used for rapidly detecting the target object in the automatic driving process.

Description

Target detection method, apparatus and computer-readable storage medium for automatic driving
Technical Field
The present application relates to the field of intelligent driving technologies, and in particular, to a method and an apparatus for detecting an object in automatic driving, and a computer-readable storage medium.
Background
Target detection is an important issue in the field of automatic driving, and in the related art, detection of a target is performed by combining image data acquired by a visual device (for example, a monocular camera, a binocular camera, a depth camera, or the like) with point cloud data acquired by a radar, that is, by fusing data acquired by two devices. However, since the data volume collected by the radar is usually large, which means a large consumption of computing resources, in the field of automatic driving, the speed of target detection is also an important indicator, because it relates to the real-time performance of the system.
Disclosure of Invention
To solve or partially solve the problems in the related art, the present application provides an object detection method, apparatus, and computer-readable storage medium for automatic driving to quickly detect an object during automatic driving.
The first aspect of the present application provides an automatic driving target detection method, including:
acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, wherein the driving environment image at least comprises an image of the surrounding environment when an intelligent vehicle drives;
under the condition that a target object in any one of at least one driving environment image or the three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object to obtain a projected area of interest;
and detecting the target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interesting area and has a range larger than that of the projection interesting area.
A second aspect of the present application provides an automatic driving target detection device, including:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, and the driving environment image at least comprises an image of the surrounding environment when an intelligent vehicle drives;
the projection module is used for projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object under the condition that the target object in any one of the driving environment image or the three-dimensional point cloud image is identified to obtain a projected area of interest;
and the calculation module is used for detecting the target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interest area, and the range of the preset target detection area is larger than that of the projection interest area.
A third aspect of the present application provides an electronic device comprising:
a processor; and
a memory having executable code stored thereon, which when executed by the processor, causes the processor to perform the method as described above.
A fourth aspect of the present application provides a computer-readable storage medium having stored thereon executable code, which, when executed by a processor of an electronic device, causes the processor to perform the method as described above.
The technical scheme provided by the application can comprise the following beneficial effects: after the projection interesting area is obtained by processing the target object, the target object is detected in a preset target detection range, and the pose information of the target object in a world coordinate system is calculated. Because it predetermines to surround the region of interest of projection and the scope is greater than the region of interest of projection to predetermine the target detection, consequently, compare the target detection real-time that prior art needs a large amount of calculations to lead to relatively poor, the technical scheme of this application only retrieves in predetermineeing the target detection region, not only ensures to detect the target object, can detect the target object fast moreover, has promoted the target detection's of autopilot real-time.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
The foregoing and other objects, features and advantages of the application will be apparent from the following more particular descriptions of exemplary embodiments of the application as illustrated in the accompanying drawings wherein like reference numbers generally represent like parts throughout the exemplary embodiments of the application.
FIG. 1 is a schematic flow chart diagram illustrating a method for detecting targets for automatic driving according to an embodiment of the present application;
fig. 2 is a schematic diagram of a bounding box of a cube surrounding a dog constructed when the object is a dog according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a projected region of interest, which is a rectangular frame projected by the square bounding box of the example in FIG. 2 according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a two-dimensional box, a square surrounding a dog, constructed when the object is the dog according to the embodiment of the present application;
FIG. 5 is a schematic diagram of a projected region of interest of a cone projected by the two-dimensional frame illustrated in FIG. 4 according to an embodiment of the present application;
FIG. 6 is a schematic diagram illustrating an enlarged square projected region of interest illustrated in FIG. 3 according to an embodiment of the present application;
FIG. 7 is a schematic diagram illustrating an enlarged cone projection region of interest illustrated in FIG. 5 according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of an object detection device for automatic driving according to an embodiment of the present application;
fig. 9 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.
Detailed Description
Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While embodiments of the present application are illustrated in the accompanying drawings, it should be understood that the present application may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this application and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
It should be understood that although the terms "first," "second," "third," etc. may be used herein to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the present application. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.
In the field of automatic driving, although the related art combines image data acquired by a visual device with point cloud data acquired by a radar, that is, combines data acquired by two devices to detect a target, the target can be detected relatively accurately, because the data volume acquired by the radar is usually large, when the point cloud data is calculated, a large amount of computing resources are consumed, and in the field of automatic driving, the speed of target detection is also an important index, so that the real-time performance of the related art is poor.
In view of the above problems, embodiments of the present application provide an automatic driving target detection method, which can quickly detect a target object in an automatic driving process.
The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Referring to fig. 1, a schematic flow chart of an automatic driving target detection method according to an embodiment of the present application is shown, which mainly includes steps S101 to S103, and is described as follows:
step S101: and acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, wherein the driving environment image at least comprises an image of the surrounding environment when the intelligent vehicle drives.
In the embodiment of the application, the camera can be a monocular camera, a binocular camera or a depth camera and other visual equipment, and both the camera and the laser radar can be carried on the intelligent vehicle. The cameras and the laser radars are good and bad in performance, for example, the running environment images collected by the cameras can easily identify information such as the types of targets in the running environment images, but are not sensitive to the position and size information of the targets; the laser radar has the advantages of strong anti-interference capability, high resolution and high ranging accuracy, but has poor applicability in severe weather such as rain, fog and snow, and the like, and the data is output in a point cloud format, so that the calculated amount is large.
Step S102: and under the condition that a target object in any one of at least one driving environment image or three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object to obtain a projected region of interest.
The identification of the target object in any one of the at least one driving environment image or the three-dimensional point cloud image means that the driving environment image acquired by the camera or the three-dimensional point cloud image obtained by scanning the laser radar is acquired, and the target object is detected in the driving environment image or the three-dimensional point cloud image through a preset algorithm. In order to finally obtain detailed information such as the type, position, orientation and the like of the target object, in the embodiment of the present application, when the target object in any one of the at least one driving environment image or the three-dimensional point cloud image is identified, a region surrounding the target object may be projected into the driving environment image or the three-dimensional point cloud image in which the target object is not identified, so as to obtain a projected region of interest.
As an embodiment of the present application, in a case that a target object in any one of at least one driving environment image or three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or three-dimensional point cloud image in which the target object is not identified, and obtaining a projected area of interest may be: when the target object in the three-dimensional point cloud image is identified, a three-dimensional region surrounding the target object is projected to the driving environment image to obtain a projected region of interest, which may be specifically implemented by the following steps S1021 to S1023:
step S1021: and acquiring a three-dimensional coordinate of each cloud point in the point cloud corresponding to the contour of the target object under a laser radar coordinate system, and constructing a three-dimensional bounding box surrounding the target object, wherein the laser radar coordinate system is a coordinate system established based on the laser radar.
And calculating to obtain the three-dimensional coordinates of each point cloud in the point cloud corresponding to the contour of the target object under the laser radar coordinate system according to the reflectivity of the laser radar, the depth measured after the laser radar strikes the target object and other information. Since the edges or the outlines of the target objects correspond to the envelopes of the target objects, a three-dimensional bounding box surrounding the target objects can be constructed as long as the three-dimensional coordinates of the cloud points corresponding to the edges or the outlines in the laser radar coordinate system are obtained, for example, a cuboid, a cube or a sphere circumscribing the target objects. In principle, any geometric body that can enclose the object can be used as a three-dimensional bounding box. Fig. 2 is a schematic diagram of a bounding box, which is a cube constructed to surround a dog when the object is the dog.
Step S1022: and acquiring a projection transformation matrix between a laser radar coordinate system and a camera coordinate system, wherein the camera coordinate system is a coordinate system established based on the camera.
Step S1023: and calculating the two-dimensional coordinates of each cloud point in the three-dimensional bounding box under the camera coordinate system based on the projection conversion matrix and the three-dimensional bounding box to obtain a projection region of interest.
Recording a projection conversion matrix between a laser radar coordinate system and a camera coordinate system as R, and assuming that the three-dimensional coordinate of any cloud point p in the three-dimensional bounding box under the laser radar coordinate system is (x)p，yp，zp) And then two-dimensional coordinates (x ') of any cloud point p in the three-dimensional bounding box under the camera coordinate system'p，y′p) Can be calculated according to the projection formula as follows:
Figure BDA0003410447160000051
when each cloud point in the three-dimensional bounding box calculates the two-dimensional coordinates of the cloud point p in the camera coordinate system according to the method for calculating the two-dimensional coordinates of the cloud point p in the camera coordinate system, the projection region of interest can be obtained. Obviously, the projected region of interest projected by the three-dimensional bounding box becomes a two-dimensional graph. Fig. 3 is a schematic diagram of a rectangular frame projected by the square bounding box of fig. 2.
In order to further reduce the calculation amount, cloud points obviously not belonging to the point cloud corresponding to the target object may be removed, in other words, for the above embodiment, the upper height limit value and the lower height limit value of the target object in the driving environment image may be obtained; determining the wild cloud points with the height values smaller than the lower height limit value or larger than the upper height limit value in the point clouds matched with the target object; the wild cloud points are filtered out from the point cloud matched with the target object, and the wild cloud points are obviously not possible to be cloud points in the point cloud corresponding to the target object, so that the filtering can be carried out in advance to reduce the calculation amount in the subsequent calculation process.
As another embodiment of the present application, in a case that a target object in any one of at least one driving environment image or three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or three-dimensional point cloud image in which the target object is not identified, and obtaining a projected area of interest may be: when a target object in the driving environment image is identified, a two-dimensional region surrounding the target object is projected to the three-dimensional point cloud image to obtain a projected region of interest, which can be specifically realized by the following steps S '1021 to S' 1023:
step S' 1021: and acquiring two-dimensional coordinates of the contour of the target object under a camera coordinate system, and constructing a two-dimensional frame surrounding the target object, wherein the camera coordinate system is a coordinate system established based on the camera.
In the embodiment of the present application, the two-dimensional frame surrounding the object may be a minimum geometric figure surrounding the object, for example, a circumscribed rectangle, a square, a circle, or the like of the object, as shown in fig. 4, which is a schematic diagram of a two-dimensional frame of a square surrounding a dog constructed when the object is the dog.
Step S' 1022: and acquiring a projection transformation matrix between a camera coordinate system and a laser radar coordinate system, wherein the laser radar coordinate system is a coordinate system established based on the laser radar.
Step S' 1023: and calculating the three-dimensional coordinate of each pixel in the two-dimensional frame under the laser radar coordinate system based on the projection conversion matrix and the two-dimensional frame to obtain a projection region of interest.
Recording a projection transformation matrix between a camera coordinate system and a laser radar coordinate system as K, and assuming that a two-dimensional coordinate of any pixel q in a two-dimensional frame under the camera coordinate system is (x)q，yq) And then three-dimensional coordinates (x ') of any pixel q in the two-dimensional frame under the laser radar coordinate system'q，y′q，z′q) Can be calculated according to the projection formula as follows:
Figure BDA0003410447160000071
and when each pixel in the two-dimensional frame calculates the three-dimensional coordinate of the pixel q in the laser radar coordinate system according to the method for calculating the three-dimensional coordinate of the pixel q in the laser radar coordinate system, the projection region of interest can be obtained. It should be noted that the projected region of interest projected from the two-dimensional frame will be a cone. Fig. 5 is a schematic diagram of a projected region of interest, which is a cone projected by the two-dimensional frame illustrated in fig. 4.
In order to be able to detect the target object, in the above embodiment, a plurality of regions corresponding to the target object may also be acquired from the three-dimensional point cloud image; acquiring a two-dimensional frame which is matched with the target object and surrounds the target object from a plurality of areas corresponding to the target object; the projected region of interest is corrected based on a two-dimensional frame matching the object and surrounding the object, and the specific method of correction may be: if the projection interest areas of at least two target objects in the three-dimensional point cloud image are in the same two-dimensional frame in the driving environment image, merging the projection interest areas of the two-dimensional frames corresponding to the at least two target objects in the three-dimensional point cloud image; or if the projection interesting area of the target object in the three-dimensional point cloud image is outside the space of the three-dimensional bounding box matched with the target object in the driving environment image, correcting the projection interesting area of the two-dimensional frame corresponding to the target object in the three-dimensional point cloud image based on the two-dimensional frame matched with the target object so that the projection interesting area of the two-dimensional frame corresponding to the target object in the point cloud image is inside the three-dimensional bounding box matched with the target object and surrounding the target object.
Step S103: and detecting a target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interesting area and has a range larger than the projection interesting area. If the target detection is performed only in the projection region of interest, the target may not be detected for various reasons, and the projection region of interest is infinitely enlarged, which is equivalent to the prior art, and will increase the calculation amount. Therefore, in the embodiment of the application, after the projection interest region is obtained, the target object can be detected in the preset target detection region and the pose information of the target object in the world coordinate system is calculated, wherein the preset target detection region surrounds the projection interest region and has a range larger than the projection interest region; further, the preset target detection area and the projection region of interest differ by a preset threshold, that is, compared with the projection region of interest, the preset target detection area is only slightly increased, which is beneficial to detecting the target object and does not increase the calculation amount. As shown in fig. 6, the dashed square frame is a preset target detection area, which is obtained by increasing the projection region of interest of the square frame illustrated in fig. 3, and the dashed cone illustrated in fig. 7 is a preset target detection area, which is obtained by increasing the projection region of interest of the cone illustrated in fig. 5.
After the steps S101 to S103, the target object may be detected in the preset target detection area and pose information of the target object in the world coordinate system, that is, coordinates, orientation, and the like may be calculated. As for the specific detection and calculation method, any existing detection and calculation method may be used, and details are not described here.
As can be seen from the above-mentioned target detection method for automatic driving illustrated in fig. 1, after a projection region of interest is obtained by processing a target object, the target object is detected in a preset target detection range, and pose information of the target object in a world coordinate system is calculated. Because it predetermines to surround the region of interest of projection and the scope is greater than the region of interest of projection to predetermine the target detection, consequently, compare the target detection real-time that prior art needs a large amount of calculations to lead to relatively poor, the technical scheme of this application only retrieves in predetermineeing the target detection region, not only ensures to detect the target object, can detect the target object fast moreover, has promoted the target detection's of autopilot real-time.
Corresponding to the embodiment of the application function implementation method, the application also provides an intelligent driving target detection device, electronic equipment and a corresponding embodiment.
Fig. 8 is a schematic structural diagram of an object detection device for automatic driving according to an embodiment of the present application. For convenience of explanation, only portions related to the embodiments of the present application are shown. The apparatus illustrated in fig. 8 mainly includes an acquisition module 801, a projection module 802, and a calculation module 803, where:
the acquisition module 801 is used for acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, wherein the driving environment image at least comprises an image of the surrounding environment when the intelligent vehicle drives;
a projection module 802, configured to, when a target object in any one of the at least one driving environment image or three-dimensional point cloud image is identified, project an area surrounding the target object into the driving environment image or three-dimensional point cloud image in which the target object is not identified, so as to obtain a projected area of interest;
the calculating module 803 is configured to detect a target object in a preset target detection area and calculate pose information of the target object in a world coordinate system, where the preset target detection area surrounds the projection interest area and has a range larger than the projection interest area.
Alternatively, the projection module 802 illustrated in fig. 8 may include a first projection processing unit, configured to, when a target object in the three-dimensional point cloud image is identified, project a three-dimensional region surrounding the target object onto the driving environment image, so as to obtain a projection region of interest.
Optionally, the first projection processing unit may include a first constructing unit, a first obtaining unit, and a first calculating unit, where:
the system comprises a first construction unit, a second construction unit and a third construction unit, wherein the first construction unit is used for acquiring the three-dimensional coordinates of each cloud point in a point cloud corresponding to the outline of a target object under a laser radar coordinate system and constructing a three-dimensional bounding box which surrounds the target object, and the laser radar coordinate system is a coordinate system established based on a laser radar;
the system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is used for acquiring a projection transformation matrix between a laser radar coordinate system and a camera coordinate system, and the camera coordinate system is a coordinate system established based on a camera;
and the first calculation unit is used for calculating the two-dimensional coordinates of each cloud point in the three-dimensional bounding box under the camera coordinate system based on the projection conversion matrix and the three-dimensional bounding box to obtain a projection region of interest.
Optionally, the apparatus illustrated in fig. 8 may further include a most value obtaining module, a determining module, and a filtering module, where:
the maximum value acquisition module is used for acquiring the upper height limit value and the lower height limit value of the target object in the driving environment image;
the determining module is used for determining the field cloud points with the height values smaller than the lower height limit value or larger than the upper height limit value in the point cloud matched with the target object;
and the filtering module is used for filtering out the wild cloud points in the point cloud matched with the target object.
Alternatively, the projection module 802 illustrated in fig. 8 may include a second projection processing unit, configured to, in a case where a target object in the driving environment image is identified, project a two-dimensional region surrounding the target object onto the three-dimensional point cloud image, so as to obtain a projection region of interest.
Optionally, the second projection processing unit may include a second constructing unit, a second acquiring unit, and a second calculating unit, where:
the second construction unit is used for acquiring a two-dimensional coordinate of the outline of the target object under a camera coordinate system and constructing a two-dimensional frame surrounding the target object, wherein the camera coordinate system is a coordinate system established based on the camera;
the second acquisition unit is used for acquiring a projection conversion matrix between a camera coordinate system and a laser radar coordinate system, wherein the laser radar coordinate system is a coordinate system established based on a laser radar;
and the second calculation unit is used for calculating the three-dimensional coordinates of each pixel in the two-dimensional frame under the laser radar coordinate system based on the projection conversion matrix and the two-dimensional frame to obtain a projection region of interest.
Optionally, the apparatus illustrated in fig. 8 may further include a multi-region acquisition module, a two-dimensional frame acquisition module, and a correction module, where:
the multi-region acquisition module is used for acquiring a plurality of regions corresponding to the target object from the three-dimensional point cloud image;
the two-dimensional frame acquisition module is used for acquiring a two-dimensional frame which is matched with the target object and surrounds the target object from a plurality of areas corresponding to the target object;
and the correction module is used for correcting the projection region of interest based on a two-dimensional frame which is matched with the target object and surrounds the target object.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Fig. 9 is a schematic structural diagram of an electronic device shown in an embodiment of the present application.
Referring to fig. 9, an electronic device 900 includes a memory 910 and a processor 920.
The Processor 920 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 910 may include various types of storage units, such as system memory, Read Only Memory (ROM), and permanent storage. Wherein the ROM may store static data or instructions for the processor 920 or other modules of the computer. The persistent storage device may be a read-write storage device. The persistent storage may be a non-volatile storage device that does not lose stored instructions and data even after the computer is powered off. In some embodiments, the persistent storage device employs a mass storage device (e.g., magnetic or optical disk, flash memory) as the persistent storage device. In other embodiments, the permanent storage may be a removable storage device (e.g., floppy disk, optical drive). The system memory may be a read-write memory device or a volatile read-write memory device, such as a dynamic random access memory. The system memory may store instructions and data that some or all of the processors require at runtime. In addition, the memory 910 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic and/or optical disks, as well. In some embodiments, memory 910 may include a removable storage device that is readable and/or writable, such as a Compact Disc (CD), a digital versatile disc read only (e.g., DVD-ROM, dual layer DVD-ROM), a Blu-ray disc read only, an ultra-dense disc, a flash memory card (e.g., SD card, min SD card, Micro-SD card, etc.), a magnetic floppy disk, or the like. Computer-readable storage media do not contain carrier waves or transitory electronic signals transmitted by wireless or wired means.
The memory 910 has stored thereon executable code that, when processed by the processor 920, may cause the processor 920 to perform some or all of the methods described above.
Furthermore, the method according to the present application may also be implemented as a computer program or computer program product comprising computer program code instructions for performing some or all of the steps of the above-described method of the present application.
Alternatively, the present application may also be embodied as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) having executable code (or a computer program or computer instruction code) stored thereon, which, when executed by a processor of an electronic device (or server, etc.), causes the processor to perform part or all of the various steps of the above-described method according to the present application.
Having described embodiments of the present application, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims (10)

1. An object detection method for automatic driving, characterized in that the method comprises:
acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, wherein the driving environment image at least comprises an image of the surrounding environment when an intelligent vehicle drives;
under the condition that a target object in any one of at least one driving environment image or the three-dimensional point cloud image is identified, projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object to obtain a projected area of interest;
and detecting the target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interesting area and has a range larger than that of the projection interesting area.
2. The automatic driving target detection method according to claim 1, wherein, when a target object in at least one of the driving environment image or the three-dimensional point cloud image is identified, projecting a region surrounding the target object into the driving environment image or the three-dimensional point cloud image in which the target object is not identified to obtain a projected region of interest, comprises: and under the condition that a target object in the three-dimensional point cloud image is identified, projecting a three-dimensional region surrounding the target object to the driving environment image to obtain the projected region of interest.
3. The automatic driving target detection method according to claim 2, wherein the projecting the three-dimensional region surrounding the target object to the driving environment image to obtain the projected region of interest when the target object in the three-dimensional point cloud image is identified comprises:
acquiring a three-dimensional coordinate of each cloud point in the point cloud corresponding to the target object outline under a laser radar coordinate system, and constructing a three-dimensional bounding box surrounding the target object, wherein the laser radar coordinate system is a coordinate system established based on the laser radar;
acquiring a projection transformation matrix between the laser radar coordinate system and a camera coordinate system, wherein the camera coordinate system is a coordinate system established based on the camera;
and calculating the two-dimensional coordinates of each cloud point in the three-dimensional bounding box under the camera coordinate system based on the projection transformation matrix and the three-dimensional bounding box to obtain the projection region of interest.
4. The autonomous-driving target detection method according to claim 3, characterized in that the method further comprises:
acquiring an upper height limit value and a lower height limit value of the target object in the driving environment image;
determining a field cloud point with a height value smaller than the lower height limit value or larger than the upper height limit value in the point cloud matched with the target object;
and filtering the wild cloud points in the point cloud matched with the target object.
5. The automatic driving target detection method according to claim 1, wherein, when a target object in at least one of the driving environment image or the three-dimensional point cloud image is recognized, projecting a region surrounding the target object into the driving environment image or the three-dimensional point cloud image in which the target object is not recognized to obtain a projected region of interest, includes: and under the condition that a target object in the driving environment image is identified, projecting a two-dimensional area surrounding the target object to the three-dimensional point cloud image to obtain the projected interesting area.
6. The automatic driving target detection method according to claim 5, wherein the projecting the projected region of interest by projecting a two-dimensional region surrounding the target object onto the three-dimensional point cloud image when the target object in the driving environment image is identified comprises:
acquiring a two-dimensional coordinate of the contour of the target object under a camera coordinate system, and constructing a two-dimensional frame surrounding the target object, wherein the camera coordinate system is a coordinate system established based on the camera;
acquiring a projection conversion matrix between a camera coordinate system and a laser radar coordinate system, wherein the laser radar coordinate system is a coordinate system established based on the laser radar;
and calculating the three-dimensional coordinate of each pixel in the two-dimensional frame under the laser radar coordinate system based on the projection conversion matrix and the two-dimensional frame to obtain the projection region of interest.
7. The autonomous-driving target detection method according to claim 6, characterized in that the method further comprises:
acquiring a plurality of areas corresponding to the target object from the three-dimensional point cloud image;
acquiring a two-dimensional frame which is matched with the target object and surrounds the target object from a plurality of areas corresponding to the target object;
correcting the projected region of interest based on a two-dimensional box that matches the object and surrounds the object.
8. An object detection device for automatic driving, characterized in that the device comprises:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring a driving environment image acquired by a camera and a three-dimensional point cloud image obtained by scanning of a laser radar, and the driving environment image at least comprises an image of the surrounding environment when an intelligent vehicle drives;
the projection module is used for projecting an area surrounding the target object into the driving environment image or the three-dimensional point cloud image without identifying the target object under the condition that the target object in any one of the driving environment image or the three-dimensional point cloud image is identified to obtain a projected area of interest;
and the calculation module is used for detecting the target object in a preset target detection area and calculating pose information of the target object in a world coordinate system, wherein the preset target detection area surrounds the projection interest area, and the range of the preset target detection area is larger than that of the projection interest area.
9. An electronic device, comprising:
a processor; and
a memory having executable code stored thereon, which when executed by the processor, causes the processor to perform the method of any one of claims 1 to 7.
10. A computer readable storage medium having stored thereon executable code which, when executed by a processor of an electronic device, causes the processor to perform the method of any of claims 1 to 7.
CN202111525882.XA 2021-12-14 2021-12-14 Target detection method, apparatus and computer-readable storage medium for automatic driving Pending CN114187579A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111525882.XA CN114187579A (en) 2021-12-14 2021-12-14 Target detection method, apparatus and computer-readable storage medium for automatic driving

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111525882.XA CN114187579A (en) 2021-12-14 2021-12-14 Target detection method, apparatus and computer-readable storage medium for automatic driving

Publications (1)

Publication Number Publication Date
CN114187579A true CN114187579A (en) 2022-03-15

Family

ID=80543711

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111525882.XA Pending CN114187579A (en) 2021-12-14 2021-12-14 Target detection method, apparatus and computer-readable storage medium for automatic driving

Country Status (1)

Country Link
CN (1) CN114187579A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115035492A (en) * 2022-06-21 2022-09-09 苏州浪潮智能科技有限公司 Vehicle identification method, device, equipment and storage medium
CN115049822A (en) * 2022-05-26 2022-09-13 中国科学院半导体研究所 Three-dimensional imaging method and device
CN115661366A (en) * 2022-12-05 2023-01-31 蔚来汽车科技(安徽)有限公司 Method for constructing three-dimensional scene model and image processing device

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115049822A (en) * 2022-05-26 2022-09-13 中国科学院半导体研究所 Three-dimensional imaging method and device
CN115035492A (en) * 2022-06-21 2022-09-09 苏州浪潮智能科技有限公司 Vehicle identification method, device, equipment and storage medium
CN115035492B (en) * 2022-06-21 2024-01-23 苏州浪潮智能科技有限公司 Vehicle identification method, device, equipment and storage medium
CN115661366A (en) * 2022-12-05 2023-01-31 蔚来汽车科技(安徽)有限公司 Method for constructing three-dimensional scene model and image processing device

Similar Documents

Publication Publication Date Title
CN114187579A (en) Target detection method, apparatus and computer-readable storage medium for automatic driving
CN111712731B (en) Target detection method, target detection system and movable platform
CN106952308B (en) Method and system for determining position of moving object
CN102915444B (en) Image registration
CN111753609B (en) Target identification method and device and camera
CN110794406B (en) Multi-source sensor data fusion system and method
WO2020063093A1 (en) Method for detecting flying spot on edge of depth image, electronic device, and computer readable storage medium
CN113920487A (en) Obstacle detection processing method, device and system
US20220266835A1 (en) Monocular vision ranging method, storage medium, and monocular camera
CN114705121B (en) Vehicle pose measurement method and device, electronic equipment and storage medium
CN115327572A (en) Method for detecting obstacle in front of vehicle
CN115100654A (en) Water level identification method and device based on computer vision algorithm
CN114529884A (en) Obstacle detection processing method, device, equipment and system based on binocular camera
CN113989755A (en) Method, apparatus and computer readable storage medium for identifying an object
WO2018220824A1 (en) Image discrimination device
CN115410168A (en) Lane line data processing method, lane line data processing apparatus, and computer-readable storage medium
CN116071421A (en) Image processing method, device and computer readable storage medium
CN112801077B (en) Method for SLAM initialization of autonomous vehicles and related device
CN113408509B (en) Signboard recognition method and device for automatic driving
WO2024060209A1 (en) Method for processing point cloud, and radar
CN114018215B (en) Monocular distance measuring method, device, equipment and storage medium based on semantic segmentation
CN113538546B (en) Target detection method, device and equipment for automatic driving
CN114332130A (en) Monocular camera acquisition method and device for high-precision images
US20240212171A1 (en) Image processing apparatus
CN114994706A (en) Obstacle detection method and device and electronic equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination