CN113486837B - Automatic driving control method for low-pass obstacle - Google Patents

Automatic driving control method for low-pass obstacle Download PDF

Info

Publication number
CN113486837B
CN113486837B CN202110815596.0A CN202110815596A CN113486837B CN 113486837 B CN113486837 B CN 113486837B CN 202110815596 A CN202110815596 A CN 202110815596A CN 113486837 B CN113486837 B CN 113486837B
Authority
CN
China
Prior art keywords
image
information
obstacle
low
pass
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.)
Active
Application number
CN202110815596.0A
Other languages
Chinese (zh)
Other versions
CN113486837A (en
Inventor
黄秋生
王宏乾
金豆
杨潘
王吉宽
淳海晏
李二宁
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.)
Anhui Jianghuai Automobile Group Corp
Original Assignee
Anhui Jianghuai Automobile Group Corp
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 Anhui Jianghuai Automobile Group Corp filed Critical Anhui Jianghuai Automobile Group Corp
Priority to CN202110815596.0A priority Critical patent/CN113486837B/en
Publication of CN113486837A publication Critical patent/CN113486837A/en
Application granted granted Critical
Publication of CN113486837B publication Critical patent/CN113486837B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0011—Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • G06T2207/30261—Obstacle
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention discloses an automatic driving control method for low-pass obstacles, which is characterized in that the accurate information of target obstacles is obtained by carrying out two-stage visual recognition on image content acquired in the continuous driving process of an automatic driving vehicle, the accurate ranging result of the low-pass obstacles and the vehicle is obtained by combining a pre-constructed image ranging engine, and finally, the self-adaptive obstacle avoidance decision guided by the type of the obstacle is realized by utilizing the recognized information of the target obstacles and the accurate ranging result. The invention realizes the accurate detection of the low-pass obstacle, and can flexibly adjust driving safety measures by matching with the detection result of the low-pass obstacle, thereby achieving the purpose of reasonably avoiding or passing through the low-pass obstacle in a targeted manner.

Description

Automatic driving control method for low-pass obstacle
Technical Field
The invention relates to the field of automatic driving, in particular to an automatic driving control method aiming at low-pass obstacles.
Background
The automatic driving automobile realizes the perception of the environment through the high-precision map, the positioning system, the radar detection system and the high-definition camera. The radar detection system can realize the distance measurement of the obstacle, the high-definition camera collects images, and whether the obstacle is a person or a car or other things can be judged through machine vision identification. The radar is suitable for detecting objects with a certain height in an induction range, but the road condition of vehicle running is complex, and the radar is difficult to detect low-pass objects such as shock absorption bands, potholes, occasional masonry, road surface water and the like on the ground, wherein the low-pass obstacle refers to an obstacle occupying a smaller area in the Z direction relative to a vehicle coordinate system (can be simply understood as a shorter obstacle); the laser radar installed on the automatic driving automobile can detect the low-pass obstacle and draw the point cloud data, but is limited by the capturing angle of the target scene object, the imaging resolution of the low-pass obstacle is not too high, and the generated three-dimensional point cloud data is difficult to accurately describe the distance of the low-pass obstacle.
Thus, the prior art does not have an ideal ranging effect on low-pass obstacles. Due to the lack of accurate distance detection, it is envisioned that the trajectory planning of the domain controller of the current-stage autopilot car in the face of low-pass obstacles is simplified.
Disclosure of Invention
In view of the above, the present invention aims to provide an automatic driving control method for low-pass obstacle, so as to obtain a more accurate detection result of the low-pass obstacle, and further realize accurate avoidance and passing control.
The technical scheme adopted by the invention is as follows:
an automatic driving control method for a low-pass obstacle, comprising:
continuously receiving images of a front road acquired by a camera mounted at the front part of a vehicle during running;
performing preliminary identification on objects in the images of each frame, and judging whether suspected low-pass degree obstacles exist in the images;
if yes, recording identification information of the suspected low-pass obstacle;
based on the identification information, carrying out fine identification on the suspected low-pass obstacle in the subsequent acquired image;
when the suspected low-pass obstacle is identified as a target obstacle, acquiring type information of the target obstacle, and combining a pre-constructed image ranging engine to acquire ranging information;
information of the target obstacle according to current vehicle running information, the ranging information and one or more of the following: and determining the avoidance passing strategy aiming at the type information of the current target obstacle according to the position information and the size information.
In at least one possible implementation manner, the obtaining ranging information in combination with the pre-constructed image ranging engine includes:
and extracting the characteristics of any frame of image used for obtaining the fine recognition result, inputting the extracted characteristics into the image ranging engine, and outputting the distance prediction information of the vehicle and the target obstacle by the image ranging engine.
In at least one possible implementation manner, the training manner of the image ranging engine includes:
setting a marker for representing a target low-pass obstacle in a road in advance;
collecting an image sample containing the marker by a camera on the vehicle;
labeling color information and distance information on the object in the image sample of each frame, wherein the distance information represents the distance between the object in the image and the vehicle;
constructing a pixel vector matrix of each frame of image sample based on the labeling result;
inputting the image samples and the corresponding pixel vector matrix into an image ranging engine, and locking the image ranging engine to the marker and the distance information from each frame of image through iterative learning;
counting, training by using all image samples to obtain stable distance information and unstable distance information, and constructing an initial coordinate vector matrix;
combining the unstable distance information to construct a coordinate offset vector matrix;
performing point multiplication by using the coordinate offset vector matrix and the initial coordinate vector matrix to optimize corresponding unstable distance information;
and updating the initial coordinate vector matrix according to the optimized distance information to obtain a target coordinate vector matrix for outputting the ranging information.
In at least one possible implementation manner, the constructing a pixel vector matrix of each frame of image samples includes:
obtaining a pixel vector matrix of each frame of image according to the pixel value imaged by the camera;
each pixel point corresponds to one element in the pixel vector matrix, and the single element at least comprises a color vector and a coordinate vector; the color vector is a three-dimensional vector constructed from three color components, and the coordinate vector is a three-dimensional vector constructed from three coordinate components with respect to the origin of the vehicle coordinate system.
In at least one possible implementation manner, the statistics are trained to obtain stable distance information and unstable distance information by using all image samples, and constructing an initial coordinate vector matrix includes:
determining pixel points with the distance information statistical variance lower than a set threshold value in the image sample as trusted pixel points; determining pixel points with the distance information statistical variance higher than a set threshold value in the image sample as unreliable pixel points;
based on all the image samples, obtaining the average value of the coordinate vectors corresponding to the trusted pixel points;
based on all the image samples, performing distance information fitting on the unreliable pixel points according to the coordinate vectors of the adjacent reliable pixel points;
and constructing an initial coordinate vector matrix by utilizing the mean value and the fitted distance information.
In at least one possible implementation manner, setting a coordinate offset vector with a value smaller than 1 for an unreliable pixel point, and determining the value of a corresponding offset vector according to a strategy that the larger the statistical variance is and the smaller the offset vector is;
and constructing a coordinate offset vector matrix by using the offset vector after the value is taken.
In at least one possible implementation manner, the method further includes: if the suspected low-pass obstacle exists in the image, stopping the acceleration mode of the vehicle and entering a pre-deceleration mode until a fine recognition result is obtained.
In at least one possible implementation manner, the method further includes: before the distance measurement, it is determined whether to start the distance measurement according to the current running information and road condition information of the vehicle.
The design concept of the invention is that the accurate information of the target obstacle is obtained by carrying out two-stage visual recognition on the image content acquired in the continuous running process of the automatic driving vehicle, the accurate ranging result of the low-pass obstacle and the vehicle is obtained by combining a pre-constructed image ranging engine, and finally the self-adaptive obstacle avoidance decision guided by the type of the obstacle is realized by utilizing the recognized information of the target obstacle and the accurate ranging result. The invention realizes the accurate detection of the low-pass obstacle, and can flexibly adjust driving safety measures by matching with the detection result of the low-pass obstacle, thereby achieving the purpose of reasonably avoiding or passing through the low-pass obstacle in a targeted manner.
Drawings
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described with reference to the accompanying drawings, in which:
fig. 1 is a flowchart of an automatic driving control method for a low-pass obstacle according to an embodiment of the present invention.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the invention.
The embodiment of the present invention proposes an automatic driving control method for a low-pass obstacle, specifically, as shown in fig. 1, may include the following steps:
step S1, continuously receiving images of a front road acquired by a camera arranged at the front part of a vehicle in the driving process;
s2, carrying out preliminary identification on objects in the images of each frame, and judging whether suspected low-pass degree obstacles exist in the images;
if yes, executing step S3, and recording identification information of the suspected low-pass obstacle; the identification information may include, but is not limited to, one or more of the following: relative position information, size information, and visual characteristic information.
Step S4, based on the identification information, carrying out fine identification on the suspected low-pass obstacle in the subsequent acquired image;
s5, when the suspected low-pass obstacle is identified as a target obstacle, acquiring type information of the target obstacle, and combining a pre-constructed image ranging engine to acquire ranging information;
step S6, according to the current vehicle running information, the ranging information and the information of one or more of the following target obstacles: and determining the avoidance passing strategy aiming at the type information of the current target obstacle according to the position information and the size information.
Further, the obtaining ranging information by combining with the pre-constructed image ranging engine includes:
and extracting the characteristics of any frame of image used for obtaining the fine recognition result, inputting the extracted characteristics into the image ranging engine, and outputting the distance prediction information of the vehicle and the target obstacle by the image ranging engine.
Further, the training mode of the image ranging engine comprises the following steps:
setting a marker for representing a target low-pass obstacle in a road in advance;
collecting an image sample containing the marker by a camera on the vehicle;
labeling color information and distance information on the object in the image sample of each frame, wherein the distance information represents the distance between the object in the image and the vehicle;
constructing a pixel vector matrix of each frame of image sample based on the labeling result;
inputting the image samples and the corresponding pixel vector matrix into an image ranging engine, and locking the image ranging engine to the marker and the distance information from each frame of image through iterative learning;
counting, training by using all image samples to obtain stable distance information and unstable distance information, and constructing an initial coordinate vector matrix;
combining the unstable distance information to construct a coordinate offset vector matrix;
performing point multiplication by using the coordinate offset vector matrix and the initial coordinate vector matrix to optimize corresponding unstable distance information;
and updating the initial coordinate vector matrix according to the optimized distance information to obtain a target coordinate vector matrix for outputting the ranging information.
Further, the constructing a pixel vector matrix of each frame of image samples includes:
obtaining a pixel vector matrix of each frame of image according to the pixel value imaged by the camera;
each pixel point corresponds to one element in the pixel vector matrix, and the single element at least comprises a color vector and a coordinate vector; the color vector is a three-dimensional vector constructed from three color components, and the coordinate vector is a three-dimensional vector constructed from three coordinate components with respect to the origin of the vehicle coordinate system.
Further, the calculating training to obtain the stable distance information and the unstable distance information by using all the image samples, and constructing an initial coordinate vector matrix includes:
determining pixel points with the distance information statistical variance lower than a set threshold value in the image sample as trusted pixel points; determining pixel points with the distance information statistical variance higher than a set threshold value in the image sample as unreliable pixel points;
based on all the image samples, obtaining the average value of the coordinate vectors corresponding to the trusted pixel points;
based on all the image samples, performing distance information fitting on the unreliable pixel points according to the coordinate vectors of the adjacent reliable pixel points;
and constructing an initial coordinate vector matrix by utilizing the mean value and the fitted distance information.
Further, setting a coordinate offset vector with a value smaller than 1 for the unreliable pixel point, and determining the value of the corresponding offset vector according to a strategy that the larger the statistical variance is, the smaller the offset vector is;
and constructing a coordinate offset vector matrix by using the offset vector after the value is taken.
Further, the method further comprises: if the suspected low-pass obstacle exists in the image, stopping the acceleration mode of the vehicle and entering a pre-deceleration mode until a fine recognition result is obtained.
Further, the method further comprises: before the distance measurement, it is determined whether to start the distance measurement according to the current running information and road condition information of the vehicle.
For ease of understanding the above embodiments and their preferred versions, the following detailed description is provided herein with reference:
during the running process of the automatic driving automobile, the camera continuously captures images of the road ahead to form a planar image (called a frame image) of one frame. The camera transmits the planar image information to a domain controller of the autopilot. The domain controller makes a rough judgment on the received planar image information, that is, whether the low-pass obstacle may exist in each frame of image (the judgment criterion may be whether there is a concentrated area of color mutation in the planar image or another quick existing logic algorithm. And if the low-pass degree obstacle is judged to be possibly present, locking information (such as, but not limited to, relative positions, visual features and the like) of the suspected obstacle, calling a preset low-pass degree obstacle identification module, and finely identifying the low-pass degree obstacle which is judged to be the suspected obstacle in the front step in a plane image acquired later in the continuous driving process (the obstacle is positioned in the subsequent image by the locked suspected obstacle information). At the same time, the domain controller instructs the autonomous vehicle to no longer have a filler door action until the image recognition module completes the final recognition analysis, at which point the vehicle enters a pre-deceleration mode, such as coasting, light braking, or enters an energy recovery state.
The foregoing fine recognition of the low-pass degree obstacles can be specifically realized based on a convolutional neural network, for example, each low-pass degree obstacle has a plurality of convolution kernels (which can be the characteristics of color, shape, water wave on the surface of accumulated water, etc.), the convolution kernels can be continuously optimized through deep learning of the network, and during actual recognition operation, the convolution kernels are utilized to check suspected obstacles in subsequent plane images to carry out multilevel pooling and activation, so that the recognition of the images is finally completed. In actual operation, the existing mature technology in the machine vision field can be adopted, and the invention is not limited to the identification technology.
If the detection result finally judges that the obstacle is not the low-pass obstacle, the identification result is returned to the domain controller, and the domain controller resumes normal automatic driving control; if the object obstacle is determined, the type of the object obstacle is determined, the recognition result is returned to the domain controller, and the domain controller calls a pre-trained image ranging engine (a neural network or other image processing models can be used for training based on a machine learning mode) to perform distance measurement on the object obstacle under the continuous driving condition (only one frame of image for obtaining the fine recognition result is needed to be used as input). It should be noted that, on the premise that the ranging algorithm of the present invention is triggered, the vehicle may travel on an open road surface with relatively high speed and relatively flat overall, rather than on a narrow road surface with relatively dense obstacles, because, in general, when traveling on a special road condition such as a narrow road surface with dense obstacles, the preset speed strategy of the autonomous vehicle generally travels at a relatively low speed due to the safety priority mechanism, it is not necessary to execute the low-pass obstacle ranging algorithm of the present invention, and it is not necessary to reduce the speed or change the traveling direction of the vehicle again.
As will be appreciated by those skilled in the art, the foregoing combination with the image ranging engine to obtain the ranging information may refer to, in actual operation, feature extraction of any frame of image for obtaining the fine recognition result, input to the image ranging engine, and output, by the image ranging engine, the distance prediction information of the vehicle and the target obstacle.
Specifically, for the training manner of the image ranging engine, the following examples may be referred to:
the method comprises the steps of arranging a marker (corresponding to a target low-pass obstacle) with a marked color mark (in other embodiments, other special mark forms can be adopted) on a relatively wide road, capturing a large number of front road images containing the marker through a camera on a running vehicle, marking each frame of image (color information can be marked), and recording the coordinate position (distance information) of an object contained in each frame of image relative to a vehicle coordinate system.
Then, converting the pixel information on each frame of image sample into an MxN vector matrix, wherein MxN is preferably a pixel imaged by a camera, namely each pixel point corresponds to one element in the vector matrix, and the element consists of a group of vectors; wherein the vector group in a single element comprises at least two types of vectors, namely a color vector alpha and a coordinate vector beta. In actual operation, the color vector α may be a three-dimensional vector composed of three components of R red, G green, and B blue; the coordinate vector β may also be a three-dimensional vector composed of X, Y, Z three coordinate components with respect to the origin of the vehicle coordinate system. Thus the foregoing 3×3 pixel vector matrix may be exemplified as follows:
α ij ＝(R ij ,G ij ,B ij ),β ij ＝(X ij ,Y ij ,Z ij )
in the training stage, a large number of image samples and the corresponding pixel vector matrixes are input into an image ranging engine, so that the image ranging engine can be trained to lock the image samples to the marker according to color information in each frame of image, and record row and column numbers (namely, which rows and columns of elements in the M multiplied by N matrix form the marker) of the marked pixel points in each frame of image, thereby obtaining coordinate vectors (the distance between the marker and a vehicle) corresponding to the pixel points where the marker is located, and obtaining the corresponding coordinate vectors from the positions where other pixel points in the image are located.
Then, relatively stable (e.g., statistical variance is lower than a certain set threshold) pixels in the image can be determined as trusted pixels according to the counted coordinate vector data, otherwise, pixels with relatively large fluctuation (e.g., statistical variance is higher than a certain set threshold) of the coordinate vector data are determined as untrusted pixels.
Then, for the trusted pixel points, obtaining corresponding coordinate vector average values based on the large amount of sample data; and fitting the unreliable pixel points according to the coordinate vectors of the adjacent reliable pixel points. From this process, an m×n initial coordinate vector matrix may be formed, in which each element represents a coordinate vector. An example of a 3 x 3 initial coordinate vector matrix is as follows:
β ij ＝(X ij ,Y ij ,Z ij )
in addition, a coordinate offset vector matrix can be constructed: in the foregoing process, the pixel point with larger fluctuation of the coordinate vector data appears, which can be considered that the image ranging engine has larger fluctuation on the distance value identified by the pixel point, the offset vector at the pixel point can be a value lower than 1 (the offset value corresponding to the stable coordinate can be 1), and the determination principle of the offset vector value can be that the larger the variance is, the smaller the offset vector value is; thus, a coordinate offset vector matrix can be constructed, which can also be an mxn matrix, with an offset vector, for example denoted as delta= (δx, δy, δz), in each element, where the three element values can be values less than 1. An example of a 3 x 3 coordinate offset vector matrix is as follows:
Δ ij ＝(δx ij ,δy ij ,δz ij )，δx ij 、δy ij 、δz ij ∈(0,1]
multiplying the offset vector matrix with the corresponding element of the initial coordinate vector matrix to form an optimized coordinate vector matrix. Specifically, for a pixel point whose coordinates fluctuate greatly, the coordinate vector thereof can be regarded as an unreliable coordinate vector, so that the coordinate vector of the pixel point is narrowed to some extent by using the offset vector, that is, the coordinate vector value at the pixel point is brought closer to the origin of the vehicle coordinate system. The purpose of the above design is that, in an actual scene, for the pixel point position where the unreliable coordinate vector is located, the expected image ranging engine can determine that the pixel point is closer to the vehicle than the content displayed by the collected current frame image, so that a certain safety margin is reserved for the subsequent countermeasure of the automatic driving vehicle.
For example, during training, if beta 23 If there is a large fluctuation in the y-coordinate in the coordinate vector at the pixel point position (the statistical variance exceeds a certain threshold value), then beta 23 Is identified as an untrusted vector; thus, delta can be set in the coordinate offset vector matrix 23 Set to (1,0.95,1). Beta 23 And delta 23 After the two vectors are subjected to dot multiplication, the generated optimized result beta' 23 And replacing the corresponding elements in the initial coordinate vector matrix, thereby realizing the optimization of the ranging result.
Therefore, the image ranging engine can directly output the accurate distance information between the vehicle and the target obstacle by using only one frame of image of the target obstacle which is finely identified as input.
After the information such as the position, the type and the distance between the target obstacle and the vehicle is obtained, the domain controller can plan the driving mode of the vehicle at the next moment so as to make the automatic driving vehicle make a corresponding decision.
For example, when the image recognition result feeds back that the target obstacle is a low-pass obstacle type such as a deceleration strip, etc., the vehicle can be controlled to enter an active deceleration mode, specifically, according to a default passing speed V0, a current vehicle speed V, a distance S between the low-pass obstacle of the type and the vehicle, a deceleration is calculated to complete deceleration passing.
When the image recognition result feeds back that the target obstacle is of a low-pass obstacle type such as relatively small masonry, a falling object of a vehicle, a well cover, a shallow pit and the like, the vehicle can be controlled to select steering avoidance according to the current speed V, the distance S between the low-pass obstacle of the type and the vehicle and the relative position, or the running direction of the vehicle can be finely adjusted, so that the low-pass obstacle is close to the center line of the vehicle, namely, the low-pass obstacle passes between the left wheel and the right wheel of the vehicle.
When the image recognition result feeds back that the target obstacle is accumulated water with a smaller area, the speed reduction passing strategy or the rarely passing strategy can be referred to, and details are not repeated here; when the target obstacle is a large-area ponding or similar low-pass obstacle, an emergency braking strategy or a gentle braking strategy can be executed according to the ranging information, so that the automatic driving automobile can be braked and parked before being involved, and a new driving route can be planned again.
In summary, the design concept of the invention is that, through performing two-stage visual recognition on the image content acquired in the continuous running process of the automatic driving vehicle, the accurate information of the target obstacle is obtained, and the accurate ranging result of the low-pass obstacle and the vehicle is obtained by combining with the image ranging engine constructed in advance, and finally, the self-adaptive obstacle avoidance decision guided by the obstacle type is realized by utilizing the recognized target obstacle information and the accurate ranging result. The invention realizes the accurate detection of the low-pass obstacle, and can flexibly adjust driving safety measures by matching with the detection result of the low-pass obstacle, thereby achieving the purpose of reasonably avoiding or passing through the low-pass obstacle in a targeted manner.
In the embodiments of the present invention, "at least one" means one or more, and "a plurality" means two or more. "and/or", describes an association relation of association objects, and indicates that there may be three kinds of relations, for example, a and/or B, and may indicate that a alone exists, a and B together, and B alone exists. Wherein A, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of the following" and the like means any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c may represent: a, b, c, a and b, a and c, b and c or a and b and c, wherein a, b and c can be single or multiple.
The construction, features and effects of the present invention are described in detail according to the embodiments shown in the drawings, but the above is only a preferred embodiment of the present invention, and it should be understood that the technical features of the above embodiment and the preferred mode thereof can be reasonably combined and matched into various equivalent schemes by those skilled in the art without departing from or changing the design concept and technical effects of the present invention; therefore, the invention is not limited to the embodiments shown in the drawings, but is intended to be within the scope of the invention as long as changes made in the concept of the invention or modifications to the equivalent embodiments do not depart from the spirit of the invention as covered by the specification and drawings.

Claims (5)

1. An automatic driving control method for a low-pass obstacle, comprising:
continuously receiving images of a front road acquired by a camera mounted at the front part of a vehicle during running;
performing preliminary identification on objects in the images of each frame, and judging whether suspected low-pass degree obstacles exist in the images;
if yes, recording identification information of the suspected low-pass obstacle;
based on the identification information, carrying out fine identification on the suspected low-pass obstacle in the subsequent acquired image;
when the suspected low-pass obstacle is identified as a target obstacle, acquiring type information of the target obstacle, and combining a pre-constructed image ranging engine to acquire ranging information;
information of the target obstacle according to current vehicle running information, the ranging information and one or more of the following: position information and size information, determining an avoidance passing strategy aiming at the type information of the current target obstacle;
the training mode of the image ranging engine comprises the following steps:
setting a marker for representing a target low-pass obstacle in a road in advance;
collecting an image sample containing the marker by a camera on the vehicle;
labeling color information and distance information on the object in the image sample of each frame, wherein the distance information represents the distance between the object in the image and the vehicle;
constructing a pixel vector matrix of each frame of image sample based on the labeling result, wherein the pixel vector matrix of each frame of image is obtained according to the pixel value imaged by the camera; each pixel point corresponds to one element in the pixel vector matrix, and the single element at least comprises a color vector and a coordinate vector; the color vector is a three-dimensional vector constructed from three color components, and the coordinate vector is a three-dimensional vector constructed from three coordinate components with respect to the origin of the vehicle coordinate system;
inputting the image samples and the corresponding pixel vector matrix into an image ranging engine, and locking the image ranging engine to the marker and the distance information from each frame of image through iterative learning;
the statistics is trained by using all image samples to obtain stable distance information and unstable distance information, and an initial coordinate vector matrix is constructed, comprising: determining pixel points with the distance information statistical variance lower than a set threshold value in the image sample as trusted pixel points; determining pixel points with the distance information statistical variance higher than a set threshold value in the image sample as unreliable pixel points; based on all the image samples, obtaining the average value of the coordinate vectors corresponding to the trusted pixel points; based on all the image samples, performing distance information fitting on the unreliable pixel points according to the coordinate vectors of the adjacent reliable pixel points; constructing an initial coordinate vector matrix by utilizing the mean value and the fitted distance information;
combining the unstable distance information to construct a coordinate offset vector matrix;
performing point multiplication by using the coordinate offset vector matrix and the initial coordinate vector matrix to optimize corresponding unstable distance information;
and updating the initial coordinate vector matrix according to the optimized distance information to obtain a target coordinate vector matrix for outputting the ranging information.
2. The method of claim 1, wherein the obtaining ranging information in combination with a pre-built image ranging engine comprises:
and extracting the characteristics of any frame of image used for obtaining the fine recognition result, inputting the extracted characteristics into the image ranging engine, and outputting the distance prediction information of the vehicle and the target obstacle by the image ranging engine.
3. The automatic driving control method for low-pass obstacles according to claim 1, wherein a coordinate offset vector with a value smaller than 1 is set for the unreliable pixel points, and the numerical value of the corresponding offset vector is determined according to a strategy that the larger the statistical variance is, the smaller the offset vector is;
and constructing a coordinate offset vector matrix by using the offset vector after the value is taken.
4. A method of autopilot control for a low-pass obstacle according to any one of claims 1 to 3, further comprising: if the suspected low-pass obstacle exists in the image, stopping the acceleration mode of the vehicle and entering a pre-deceleration mode until a fine recognition result is obtained.
5. A method of autopilot control for a low-pass obstacle according to any one of claims 1 to 3, further comprising: before the distance measurement, it is determined whether to start the distance measurement according to the current running information and road condition information of the vehicle.
CN202110815596.0A 2021-07-19 2021-07-19 Automatic driving control method for low-pass obstacle Active CN113486837B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110815596.0A CN113486837B (en) 2021-07-19 2021-07-19 Automatic driving control method for low-pass obstacle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110815596.0A CN113486837B (en) 2021-07-19 2021-07-19 Automatic driving control method for low-pass obstacle

Publications (2)

Publication Number Publication Date
CN113486837A CN113486837A (en) 2021-10-08
CN113486837B true CN113486837B (en) 2023-07-18

Family

ID=77941448

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110815596.0A Active CN113486837B (en) 2021-07-19 2021-07-19 Automatic driving control method for low-pass obstacle

Country Status (1)

Country Link
CN (1) CN113486837B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114407901B (en) * 2022-02-18 2023-12-19 北京小马易行科技有限公司 Control method and device for automatic driving vehicle and automatic driving system

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4407920B2 (en) * 2004-05-19 2010-02-03 ダイハツ工業株式会社 Obstacle recognition method and obstacle recognition device
JP4407921B2 (en) * 2004-05-19 2010-02-03 ダイハツ工業株式会社 Obstacle recognition method and obstacle recognition device
JP2008298533A (en) * 2007-05-30 2008-12-11 Konica Minolta Holdings Inc Obstruction measurement method, device, and system
JP2015194798A (en) * 2014-03-31 2015-11-05 日産自動車株式会社 Driving assistance control device
JP6610585B2 (en) * 2017-03-13 2019-11-27 トヨタ自動車株式会社 Collision avoidance control device
CN109116374B (en) * 2017-06-23 2021-08-17 百度在线网络技术（北京）有限公司 Method, device and equipment for determining distance of obstacle and storage medium
CN108596058A (en) * 2018-04-11 2018-09-28 西安电子科技大学 Running disorder object distance measuring method based on computer vision
CN110909569B (en) * 2018-09-17 2022-09-23 深圳市优必选科技有限公司 Road condition information identification method and terminal equipment
CN110147706B (en) * 2018-10-24 2022-04-12 腾讯科技（深圳）有限公司 Obstacle recognition method and device, storage medium, and electronic device
CN113454636A (en) * 2018-12-28 2021-09-28 辉达公司 Distance of obstacle detection in autonomous machine applications
CN109829403B (en) * 2019-01-22 2020-10-16 淮阴工学院 Vehicle anti-collision early warning method and system based on deep learning
CN110688903B (en) * 2019-08-30 2023-09-26 湖南九域同创高分子新材料有限责任公司 Barrier extraction method based on train AEB system camera data
CN110825093B (en) * 2019-11-28 2021-04-16 安徽江淮汽车集团股份有限公司 Automatic driving strategy generation method, device, equipment and storage medium
CN111046843B (en) * 2019-12-27 2023-06-20 华南理工大学 Monocular ranging method in intelligent driving environment
CN111832418A (en) * 2020-06-16 2020-10-27 北京汽车研究总院有限公司 Vehicle control method, device, vehicle and storage medium
CN112014845B (en) * 2020-08-28 2024-01-30 安徽江淮汽车集团股份有限公司 Vehicle obstacle positioning method, device, equipment and storage medium
CN112180951A (en) * 2020-11-10 2021-01-05 桃江县缘湘聚文化传媒有限责任公司 Intelligent obstacle avoidance method for unmanned vehicle and computer readable storage medium
CN112373467A (en) * 2020-11-10 2021-02-19 桃江县缘湘聚文化传媒有限责任公司 Intelligent obstacle avoidance system of unmanned automobile

Also Published As

Publication number Publication date
CN113486837A (en) 2021-10-08

Similar Documents

Publication Publication Date Title
US11093801B2 (en) Object detection device and object detection method
CN109460709B (en) RTG visual barrier detection method based on RGB and D information fusion
CN110765922B (en) Binocular vision object detection obstacle system for AGV
CN104573646B (en) Chinese herbaceous peony pedestrian detection method and system based on laser radar and binocular camera
CN109334563B (en) Anti-collision early warning method based on pedestrians and riders in front of road
CN108596058A (en) Running disorder object distance measuring method based on computer vision
RU2568777C2 (en) Device to detect moving bodies and system to detect moving bodies
CN107972662A (en) To anti-collision warning method before a kind of vehicle based on deep learning
CN108243623A (en) Vehicle anticollision method for early warning and system based on binocular stereo vision
EP2960858B1 (en) Sensor system for determining distance information based on stereoscopic images
CN111891061B (en) Vehicle collision detection method and device and computer equipment
CN108108750A (en) Metric space method for reconstructing based on deep learning and monocular vision
JP7135665B2 (en) VEHICLE CONTROL SYSTEM, VEHICLE CONTROL METHOD AND COMPUTER PROGRAM
CN110717445A (en) Front vehicle distance tracking system and method for automatic driving
CN115331191B (en) Vehicle type recognition method, device, system and storage medium
CN114419098A (en) Moving target trajectory prediction method and device based on visual transformation
CN104951758A (en) Vehicle-mounted method and vehicle-mounted system for detecting and tracking pedestrians based on vision under urban environment
CN113435237A (en) Object state recognition device, recognition method, recognition program, and control device
CN113486837B (en) Automatic driving control method for low-pass obstacle
CN111079675A (en) Driving behavior analysis method based on target detection and target tracking
CN112115800A (en) Vehicle combination recognition system and method based on deep learning target detection
CN113221739B (en) Monocular vision-based vehicle distance measuring method
JP4967758B2 (en) Object movement detection method and detection apparatus
JP7226368B2 (en) Object state identification device
CN111414857A (en) Front vehicle detection method based on vision multi-feature fusion

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
CB02 Change of applicant information

Address after: No.669 Shixin Road, economic development zone, Feixi County, Hefei City, Anhui Province

Applicant after: ANHUI JIANGHUAI AUTOMOBILE GROUP Corp.,Ltd.

Address before: 230601 No. 669 Shixin Road, Taohua Industrial Park, Hefei City, Anhui Province

Applicant before: ANHUI JIANGHUAI AUTOMOBILE GROUP Corp.,Ltd.

GR01 Patent grant
GR01 Patent grant