CN111695373A - Zebra crossing positioning method, system, medium and device - Google Patents
Zebra crossing positioning method, system, medium and device Download PDFInfo
- Publication number
- CN111695373A CN111695373A CN201910185277.9A CN201910185277A CN111695373A CN 111695373 A CN111695373 A CN 111695373A CN 201910185277 A CN201910185277 A CN 201910185277A CN 111695373 A CN111695373 A CN 111695373A
- Authority
- CN
- China
- Prior art keywords
- image
- zebra crossing
- zebra
- rectangular frame
- monitoring image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 241000283070 Equus zebra Species 0.000 title claims abstract description 193
- 238000000034 method Methods 0.000 title claims abstract description 69
- 238000012544 monitoring process Methods 0.000 claims abstract description 60
- 238000001514 detection method Methods 0.000 claims abstract description 46
- 238000012545 processing Methods 0.000 claims abstract description 36
- 238000013135 deep learning Methods 0.000 claims abstract description 32
- 238000006243 chemical reaction Methods 0.000 claims abstract description 21
- 238000004458 analytical method Methods 0.000 claims abstract description 19
- 238000007621 cluster analysis Methods 0.000 claims abstract description 19
- 238000003708 edge detection Methods 0.000 claims abstract description 13
- 230000005484 gravity Effects 0.000 claims description 27
- 238000004590 computer program Methods 0.000 claims description 11
- 230000001186 cumulative effect Effects 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 description 6
- 238000013527 convolutional neural network Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method, a system, a medium and equipment for positioning a zebra crossing, wherein the method for positioning the zebra crossing comprises the following steps: detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarization image of the monitoring image; white pixel points in the binary image are subjected to clustering analysis, and noise areas of non-zebra stripes are removed; and selecting contour extreme points in the binary image after the cluster analysis through edge detection, and extracting zebra stripes to locate zebra stripe areas. The positioning method of the zebra crossing provided by the invention improves the accuracy of positioning the zebra crossing and provides a basis for complex vehicle running analysis at a traffic intersection.
Description
Technical Field
The invention belongs to the field of image processing, relates to a zebra crossing image positioning method, and particularly relates to a zebra crossing positioning method, a zebra crossing positioning system, a zebra crossing positioning medium and zebra crossing positioning equipment.
Background
The problem of zebra crossing detection is that automatic driving and intelligent traffic are always concerned about, and the zebra crossing shape and size are different due to the difference of the angles of the cameras. Under the condition of automatic driving, the zebra stripes are mostly parallel to the upper and lower boundaries in the collected images, and the traffic cameras, especially the cameras capable of rotating angles, have different directions, and under the scene of a traffic intersection, a plurality of clear or fuzzy zebra stripes which are greatly deformed due to the perspective effect exist in one picture. This situation presents additional difficulties for the identification and location of zebra stripes.
The problem of zebra crossing detection is compatible with the development of automatic driving, and is not limited to automatic driving scenes, for example, the problem that detection and positioning of zebra crossings in intelligent traffic must be overcome. The traditional zebra crossing detection method focuses on the physical characteristics of zebra crossings, extracts edges through a canny operator, obtains straight lines through Hough transform, and detects the existence of the zebra crossings through parallel analysis of a series of straight lines. However, the methods are different in magnitude, and have poor adaptability to zebra crossing detection under different light, deformation and definition scenes.
Deep learning is a popular direction in recent years for computer and even other disciplinary research. Deep learning frameworks such as deep neural networks, convolutional neural networks, deep belief networks, recurrent neural networks and the like are widely applied to the fields of computer vision, speech recognition, natural language processing, audio recognition, bioinformatics and the like, and have a good effect. The benefit of deep learning is to efficiently replace manually acquired features with semi-supervised or unsupervised feature learning and hierarchical feature extraction.
In the deep learning target detection algorithm, the Faster R-CNN is an algorithm with high speed and high accuracy, integrates four basic steps (candidate region generation, feature extraction, classification and position adjustment) of target detection, and greatly improves the running speed. However, in the conventional target detection algorithm, the generated candidate area and the detection frame output by the network are horizontal rectangular frames, and a good positioning effect cannot be achieved for non-conventional target detection data (such as a long and narrow zebra crossing at a traffic intersection).
Therefore, how to provide a method for positioning a zebra crossing area, so that the positioning is more accurate and more suitable for a complicated and variable environment to reduce the false detection rate of the zebra crossing, has become a technical problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the above drawbacks of the prior art, an object of the present invention is to provide a method, a system, a medium, and a device for locating a zebra crossing, which are used to solve the problem of low accuracy in locating the zebra crossing area in the prior art.
In order to achieve the above and other related objects, the present invention provides a zebra crossing locating method, including: detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarization image of the monitoring image; white pixel points in the binary image are subjected to clustering analysis, and noise areas of non-zebra stripes are removed; and selecting contour extreme points in the binary image after the cluster analysis through edge detection, and extracting zebra stripes to locate zebra stripe areas.
In an embodiment of the present invention, before the step of detecting the acquired monitoring image of the traffic intersection with a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining a zebra crossing position in an image position area of each identified zebra crossing in the monitoring image, the method for positioning the zebra crossing further includes: and training a deep learning zebra crossing detection model by using the pre-collected monitoring images of a plurality of traffic intersections to generate the zebra crossing detection network.
In an embodiment of the present invention, the step of performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarized image of the monitoring image includes: performing color space conversion on the traffic monitoring image in the horizontal rectangular frame; drawing an accumulative histogram according to the proportion of the pixel points corresponding to the brightness values in the total number of the pixel points after the color space conversion; and setting a proportional threshold value in the cumulative histogram as a boundary value of a white pixel point of image binarization.
In an embodiment of the present invention, the step of performing cluster analysis on white pixel points in the binarized image and eliminating noise regions other than zebra crossings includes: in each horizontal rectangular frame region, carrying out density analysis on regions of all the white pixel points in the binary image according to a preset size parameter; determining the average position of the white pixel points in each horizontal rectangular frame area one by one according to the result of the density analysis, and taking the average position as the clustering gravity center; performing straight line fitting on all the clustering gravity centers; and eliminating the area of the white pixel point corresponding to the clustering gravity center deviated from the straight line.
In an embodiment of the present invention, the step of removing the area of the white pixel point corresponding to the cluster center deviating from the straight line includes: determining the area of the white pixel point corresponding to the clustering center deviating from the straight line as a noise area; and eliminating the noise area.
In an embodiment of the invention, the clustering barycenter is a density barycenter of the pixels determined in the horizontal rectangular frame.
In an embodiment of the invention, the cluster analysis is based on a preset scan radius parameter and a minimum inclusion point parameter.
Another aspect of the present invention provides a system for locating a zebra crossing, including: the position selection module is used for detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; the binarization processing module is used for carrying out color space conversion on the monitoring image in the horizontal rectangular frame and carrying out image binarization processing to generate a binarization image of the monitoring image; the cluster processing module is used for carrying out cluster analysis on white pixel points in the binary image and eliminating noise areas of non-zebra stripes; and the positioning module is used for selecting an extreme point of the contour from the binary image after the cluster analysis through edge detection, extracting the zebra crossing contour and positioning the zebra crossing region.
Yet another aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method for locating zebra crossing.
A final aspect of the invention provides an apparatus comprising: a processor and a memory;
the memory is used for storing a computer program, and the processor is used for executing the computer program stored by the memory so as to enable the equipment to execute the zebra crossing positioning method.
As described above, the zebra crossing positioning method, system, medium and apparatus of the present invention have the following beneficial effects:
the method can quickly and accurately identify the zebra crossing in the image acquired by the camera at the traffic intersection, can adapt to different weather, time and seasons to accurately position the zebra crossing, greatly reduces the false detection rate of the zebra crossing, improves the accuracy rate of the zebra crossing positioning, and provides a foundation for complex vehicle running analysis at the traffic intersection.
Drawings
Fig. 1 is a schematic flow chart illustrating a method for locating a zebra crossing according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a zebra crossing image detected by the zebra crossing detection network.
Fig. 3 is a schematic system diagram of a zebra crossing location system according to an embodiment of the present invention.
Description of the element reference numerals
21 horizontal rectangle frame
22 white pixel point region
23 edge detection line
24 non-zebra crossing area
25 center of gravity of cluster on fitted straight line
26 clustered centers of gravity that deviate from the fitted line
27 fitting a straight line
3 positioning system of zebra crossing
31 position selection module
32 binarization processing module
33 clustering processing module
34 positioning module
S11-S14 zebra crossing positioning method
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
The technical principles of the method, the system, the medium and the equipment for positioning the zebra crossing are as follows: detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarization image of the monitoring image; white pixel points in the binary image are subjected to clustering analysis, and noise areas of non-zebra stripes are removed; and selecting contour extreme points in the binary image after the cluster analysis through edge detection, and extracting zebra stripes to locate zebra stripe areas.
Example one
The method for positioning the zebra crossing provided by the embodiment comprises the following steps: detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarization image of the monitoring image; white pixel points in the binary image are subjected to clustering analysis, and noise areas of non-zebra stripes are removed; and selecting contour extreme points in the binary image after the cluster analysis through edge detection, and extracting zebra stripes to locate zebra stripe areas.
The method for locating a zebra crossing provided by the present embodiment will be described in detail with reference to the drawings.
Referring to fig. 1, a schematic flow chart of a zebra crossing location method according to an embodiment of the present invention is shown. As shown in fig. 1, the method for locating a zebra crossing specifically includes the following steps:
and S10, training a deep learning zebra crossing detection model by using the pre-collected monitoring images of the plurality of traffic intersections, and generating the zebra crossing detection network.
The monitoring images are images shot by cameras at actual traffic intersections, the positions of zebra stripes in the images are different, the shapes and the sizes of the zebra stripes are not similar to each other, the difference between the brightness and the darkness of the images is large according to shooting time, weather and other reasons, and the zebra stripes are large in definition difference due to different drawing time in actual scenes.
S11, detecting the collected monitoring image of the traffic intersection through a zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in the image position area of each identified zebra crossing in the monitoring image.
The zebra crossing detection network is obtained by training through a pre-stored deep learning method. The deep learning method greatly improves the detection accuracy of the traditional method, reduces the omission factor, improves the detection accuracy and improves the detection universality, and the method is more suitable for zebra crossing detection under the unfavorable environment conditions of fuzzy and dark images and the like and is also more suitable for zebra crossing detection with different shapes, and is not limited to the zebra crossing region of interest determined according to parallel edges and white regions.
In practical application, the deep learning method adopts a deep learning Faster R-CNN algorithm to detect zebra stripes appearing in an image. The detection network is trained by 400 training samples, and approximately comprises more than 1200 zebra stripes. And detecting the zebra crossing in the test picture through the trained model to obtain a horizontal rectangular frame which is used as an interest region in post-processing. Please refer to fig. 2, which shows a schematic diagram of a zebra crossing image detected by the zebra crossing detection network.
In this embodiment, the acquired monitoring image of the traffic intersection is detected by using a deep learning zebra crossing detection network, and for each identified zebra crossing in the monitoring image, a horizontal rectangular frame for determining the position of the zebra crossing is set in the image position area of the zebra crossing.
In the embodiment, zebra crossing identification is performed through a fast R-CNN deep learning target detection algorithm, the color characteristics of the zebra crossing are extracted, and the zebra crossing position is determined through a clustering algorithm and contour extraction. Due to the differences in the actual scenes, the fast R-CNN algorithm for deep learning in the step S11 plays a role in learning the zebra crossing characteristics of various different scenes and accurately identifying the zebra crossing in various real scenes, and compared with other traditional methods, the deep learning method has universality and is more inclusive.
And S12, performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarized image of the monitoring image.
In this embodiment, the S12 includes: performing color space conversion on the traffic monitoring image in the horizontal rectangular frame; drawing an accumulative histogram according to the proportion of the pixel points corresponding to the brightness values in the total number of the pixel points after the color space conversion; and setting a proportional threshold value in the cumulative histogram as a boundary value of a white pixel point of image binarization.
Specifically, the horizontal rectangular frame 21 obtained by the deep learning detection algorithm in step S11 is used as a region of interest, and image binarization is performed in the region, where the binarization method is a cumulative histogram method.
A color cumulative histogram binarization method is an adaptive threshold value method, and is characterized in that a cumulative histogram is manufactured by counting the proportion of pixel points corresponding to Y channels, namely luminance channels 0-255, in a YCbCr color space to the total number of the pixel points, and the proportion of white pixel points after binarization is artificially set to binarize an interested region.
Because the brightness of the zebra crossing in the image is different due to the difference of light rays in the natural environment, a brightest partial area in the region of interest needs to be selected by a self-adaptive method to obtain the approximate position range of the zebra crossing. In the S12 step, the color space of the RGB image is first converted into the YCbCr color space. The YCbCr color space Y is a brightness component of the color, so that compared with other binarization methods, the extraction of the region with the highest brightness in the region of interest is more favorable for segmenting the zebra crossing under different illumination conditions. The conversion formula is as follows:
compared with other image binarization methods, the binary image obtained by using the color cumulative histogram method in the embodiment is more suitable for traffic camera shooting scenes with different image brightness degrees.
In this embodiment, white areas of the binarized image are clustered, an unsupervised DBSCAN algorithm is used, clustering is performed according to density, and clustering of each white line can be performed on zebra stripes with large areas nearby, so that segmentation of the zebra stripes is more accurate, the white areas far away from the zebra stripe main body can be segmented, and positioning is more accurate.
S13, performing cluster analysis on white pixel points in the binary image, and eliminating noise areas of non-zebra stripes;
specifically, the DBSCAN algorithm is used for clustering the area 22 of the white pixel points in the binary image, the pixel points of the non-zebra stripe area 24 are eliminated, and the segmentation according to the zebra stripe is realized. The DBSCAN is a clustering analysis algorithm based on density, two parameters are needed to scan the radius parameter epsilon and the minimum point number needed for forming the high-density area, namely the minimum contained point number, and the zebra stripes in the interested area and the noise which is binary to white can be clustered respectively by selecting a certain scanning radius parameter epsilon and the minimum point number needed for forming the high-density area.
In this embodiment, the S13 includes:
a. in each horizontal rectangular frame region, carrying out density analysis on regions of all the white pixel points in the binary image according to a preset size parameter;
a suitable parameter is designed for each horizontal rectangular frame 21 with the zebra crossing detected in the step S11 according to the size for clustering, so that the zebra crossings with different sizes can be effectively subjected to adaptive clustering, and the zebra crossings and the noise can be clustered by a DBSCAN algorithm for each rectangular frame with different sizes.
b. The density analysis is to analyze the pixel point distribution position of the white pixel point region by using a DBSCAN algorithm, determine a position distribution gravity center according to the pixel point distribution position, determine the average position of the white pixel point in each horizontal rectangular frame region one by one according to the position distribution gravity center of the density analysis, and take the average position as the clustering gravity center;
and respectively calculating the clustering gravity centers of all the classes obtained by the clustering algorithm, wherein the obtained gravity centers can be regarded as the average positions of the points in the classes due to the same density of the pixel points.
c. Performing straight line fitting on all the clustering barycenters through a linear fitting algorithm;
by fitting each center point line to obtain a fitted line 27, the distance from the center of gravity of the noise such as the clustering center of gravity 26 deviating from the fitted line to the line is larger, so that the rejection is set to black, and other pixel points are reserved such as the clustering center of gravity 25 on the fitted line as the pixel points representing the zebra stripes. And the straight line fitted by the cluster center of gravity 25 on the remaining fitted straight line can represent the zebra crossing direction.
d. And eliminating the area of the white pixel point corresponding to the clustering gravity center deviated from the straight line.
In this embodiment, the step of rejecting the region of the white pixel corresponding to the cluster center deviating from the straight line includes: determining the area of the white pixel point corresponding to the clustering center deviating from the straight line as a noise area; and eliminating the noise area.
In this embodiment, the clustering center of gravity is the density center of gravity of the pixel points determined in the horizontal rectangular frame.
In this embodiment, the cluster analysis is performed according to a preset scan radius parameter and a minimum inclusion point parameter.
And S14, obtaining an edge detection line 23 in the binary image after the cluster analysis through edge detection, as shown in FIG. 2, selecting a contour extreme point A, B, C, D, and extracting a zebra crossing contour to locate a zebra crossing region.
In this embodiment, by extracting the contour of the edge detection lines 23 on both sides of the zebra crossing and selecting the contour extreme point A, B, C, D, noise interference can be eliminated, and a polygonal zebra crossing contour can be obtained.
The present embodiment also provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the method for locating a zebra crossing.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the above method embodiments may be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The method solves the problems that the universality is not strong in the traditional method and the detection frame of the deep learning method does not achieve ideal positioning accuracy, combines the advantages of the deep learning method and the traditional method, makes up the defects mutually, and achieves a relatively accurate traffic intersection zebra crossing positioning effect.
Example two
This embodiment provides a positioning system of zebra crossing, positioning system of zebra crossing includes: the position selection module is used for detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image; the binarization processing module is used for carrying out color space conversion on the monitoring image in the horizontal rectangular frame and carrying out image binarization processing to generate a binarization image of the monitoring image; the cluster processing module is used for carrying out cluster analysis on white pixel points in the binary image and eliminating noise areas of non-zebra stripes; and the positioning module is used for selecting an extreme point of the contour from the binary image after the cluster analysis through edge detection, extracting the zebra crossing contour and positioning the zebra crossing region.
The positioning system for zebra crossing provided by the present embodiment will be described in detail with reference to the drawings. The positioning system of the zebra crossing described in this embodiment is applied to the positioning method of the zebra crossing shown in fig. 1 and fig. 2.
Referring to fig. 3, fig. 3 is a schematic system diagram illustrating a zebra crossing location system according to an embodiment of the present invention. As shown in fig. 3, the zebra crossing locating system 3 includes: a position selecting module 31, a binarization processing module 32, a clustering processing module 33 and a positioning module 34.
The position selection module 31 is utilized to detect the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and a horizontal rectangular frame for determining the zebra crossing position is arranged in the image position area of each identified zebra crossing in the monitoring image;
specifically, the position selecting module 31 is configured to detect the acquired monitoring image of the traffic intersection with a deep learning zebra crossing detection network, and set a horizontal rectangular frame for determining the zebra crossing position in an image position area of each identified zebra crossing in the monitoring image.
Performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing through a binarization processing module 32 to generate a binarization image of the monitoring image;
specifically, the binarization processing module 32 is configured to perform color space conversion on the traffic monitoring image in the horizontal rectangular frame; drawing an accumulative histogram according to the proportion of the pixel points corresponding to the brightness values in the total number of the pixel points after the color space conversion; and setting a proportional threshold value in the cumulative histogram as a boundary value of a white pixel point of image binarization.
The clustering processing module 33 is used for clustering and analyzing white pixel points in the binarized image and eliminating noise regions of non-zebra stripes;
specifically, the clustering module 33 is configured to perform, in each horizontal rectangular frame region, density analysis on regions of all the white pixel points in the binarized image according to a preset size parameter; determining the average position of the white pixel points in each horizontal rectangular frame area one by one according to the result of the density analysis, and taking the average position as the clustering gravity center; performing straight line fitting on all the clustering gravity centers; and eliminating the area of the white pixel point corresponding to the clustering gravity center deviated from the straight line.
In this embodiment, a region of a white pixel point corresponding to the clustering center deviating from a straight line is determined as a noise region; and eliminating the noise area.
In this embodiment, the clustering center of gravity is the density center of gravity of the pixel points determined in the horizontal rectangular frame.
In this embodiment, the cluster analysis is performed according to a preset scan radius parameter and a minimum inclusion point parameter.
And selecting an extreme point of the contour in the binary image after the cluster analysis through edge detection by a positioning module 34, and extracting the zebra crossing contour to position the zebra crossing region.
It should be noted that the division of the modules of the positioning system is only a logical division, and the actual implementation may be wholly or partially integrated into one physical entity or may be physically separated. And these modules can be realized in the form of software called by processing element; or may be implemented entirely in hardware; and part of the modules can be realized in the form of calling software by the processing element, and part of the modules can be realized in the form of hardware. For example, the x module may be a processing element that is set up separately, or may be implemented by being integrated in a chip of the positioning system, or may be stored in a memory of the positioning system in the form of program code, and the function of the x module may be called and executed by a processing element of the positioning system. Other modules are implemented similarly. In addition, all or part of the modules can be integrated together or can be independently realized. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in the form of software.
For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or one or more microprocessors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), etc. For another example, when one of the above modules is implemented in the form of a processing element scheduler code, the processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).
The method solves the problems that the universality is not strong in the traditional method and the detection frame of the deep learning method does not achieve ideal positioning accuracy, combines the advantages of the deep learning method and the traditional method, makes up the defects mutually, and achieves a relatively accurate traffic intersection zebra crossing positioning effect.
EXAMPLE III
The present embodiment provides an apparatus, comprising: a processor, memory, transceiver, communication interface, or/and system bus; the memory is used for storing the computer program, the communication interface is used for communicating with other devices, and the processor and the transceiver are used for running the computer program to enable the devices to execute the steps of the zebra crossing positioning method according to the embodiment one.
The above-mentioned system bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The system bus may be divided into an address bus, a data bus, a control bus, and the like. The communication interface is used for realizing communication between the database access device and other equipment (such as a client, a read-write library and a read-only library). The Memory may include a Random Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, or discrete hardware components.
The protection scope of the zebra crossing location method is not limited to the execution sequence of the steps listed in this embodiment, and all the solutions implemented by adding, subtracting and replacing the steps in the prior art according to the principles of the present invention are included in the protection scope of the present invention.
The invention also provides a positioning system of the zebra crossing, which can realize the positioning method of the zebra crossing, but the device for realizing the positioning method of the zebra crossing comprises but is not limited to the structure of the positioning system of the zebra crossing listed in the embodiment, and all structural modifications and replacements in the prior art made according to the principle of the invention are included in the protection scope of the invention.
In conclusion, the method, the system, the medium and the equipment for locating the zebra crossings can quickly and accurately identify the zebra crossings in the images acquired by the cameras at the traffic intersections, can adapt to different weather, time and seasons to accurately locate the zebra crossings, greatly reduce the false detection rate of the zebra crossings, improve the accuracy rate of the zebra crossing location, provide a foundation for complex vehicle running analysis at the traffic intersections, and effectively overcome various defects in the prior art, thereby having high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (10)
1. A zebra crossing positioning method is characterized by comprising the following steps:
detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image;
performing color space conversion on the monitoring image in the horizontal rectangular frame, and performing image binarization processing to generate a binarization image of the monitoring image;
white pixel points in the binary image are subjected to clustering analysis, and noise areas of non-zebra stripes are removed;
and selecting contour extreme points in the binary image after the cluster analysis through edge detection, and extracting zebra stripes to locate zebra stripe areas.
2. The method according to claim 1, wherein before the step of detecting the acquired monitoring image of the traffic intersection with a deep learning zebra crossing detection network and setting a horizontal rectangular frame for determining the zebra crossing position in the image position area of each identified zebra crossing in the monitoring image, the method further comprises:
and training a deep learning zebra crossing detection model by using the pre-collected monitoring images of a plurality of traffic intersections to generate the zebra crossing detection network.
3. The zebra crossing positioning method according to claim 1, wherein the step of performing color space conversion on the monitoring image in the horizontal rectangular frame and performing image binarization processing to generate a binarized image of the monitoring image comprises:
performing color space conversion on the traffic monitoring image in the horizontal rectangular frame;
drawing an accumulative histogram according to the proportion of the pixel points corresponding to the brightness values in the total number of the pixel points after the color space conversion;
and setting a proportional threshold value in the cumulative histogram as a boundary value of a white pixel point of image binarization.
4. The zebra crossing positioning method according to claim 3, wherein the step of clustering white pixel points in the binarized image and eliminating noise regions other than zebra crossings comprises:
in each horizontal rectangular frame region, carrying out density analysis on regions of all the white pixel points in the binary image according to a preset size parameter;
determining the average position of the white pixel points in each horizontal rectangular frame area one by one according to the result of the density analysis, and taking the average position as the clustering gravity center;
performing straight line fitting on all the clustering gravity centers;
and eliminating the area of the white pixel point corresponding to the clustering gravity center deviated from the straight line.
5. The zebra crossing positioning method according to claim 4, wherein the step of eliminating the white pixel point region corresponding to the cluster center deviating from the straight line comprises:
determining the area of the white pixel point corresponding to the clustering center deviating from the straight line as a noise area;
and eliminating the noise area.
6. The method of claim 4, wherein the Zebra stripe is a stripe of a rectangular shape,
and the clustering gravity center is the density gravity center of the pixel points determined in the horizontal rectangular frame.
7. The method of claim 4, wherein the Zebra stripe is a stripe of a rectangular shape,
and the cluster analysis is carried out according to a preset scanning radius parameter and a minimum contained point number parameter.
8. A zebra crossing location system, comprising:
the position selection module is used for detecting the collected monitoring image of the traffic intersection by using a deep learning zebra crossing detection network, and setting a horizontal rectangular frame for determining the position of the zebra crossing in an image position area of each identified zebra crossing in the monitoring image;
the binarization processing module is used for carrying out color space conversion on the monitoring image in the horizontal rectangular frame and carrying out image binarization processing to generate a binarization image of the monitoring image;
the cluster processing module is used for carrying out cluster analysis on white pixel points in the binary image and eliminating noise areas of non-zebra stripes;
and the positioning module is used for selecting an extreme point of the contour from the binary image after the cluster analysis through edge detection, extracting the zebra crossing contour and positioning the zebra crossing region.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method for locating a zebra crossing as claimed in any one of claims 1 to 7.
10. An apparatus, comprising: a processor and a memory;
the memory is used for storing a computer program, and the processor is used for executing the computer program stored by the memory to enable the device to execute the zebra crossing positioning method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910185277.9A CN111695373B (en) | 2019-03-12 | 2019-03-12 | Zebra stripes positioning method, system, medium and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910185277.9A CN111695373B (en) | 2019-03-12 | 2019-03-12 | Zebra stripes positioning method, system, medium and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111695373A true CN111695373A (en) | 2020-09-22 |
CN111695373B CN111695373B (en) | 2024-02-02 |
Family
ID=72475678
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910185277.9A Active CN111695373B (en) | 2019-03-12 | 2019-03-12 | Zebra stripes positioning method, system, medium and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111695373B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112241714A (en) * | 2020-10-22 | 2021-01-19 | 北京字跳网络技术有限公司 | Method and device for identifying designated area in image, readable medium and electronic equipment |
CN112446315A (en) * | 2020-11-19 | 2021-03-05 | 武汉中海庭数据技术有限公司 | High-precision map zebra crossing fusion method, system, server and storage medium |
CN112633151A (en) * | 2020-12-22 | 2021-04-09 | 浙江大华技术股份有限公司 | Method, device, equipment and medium for determining zebra crossing in monitored image |
CN113435350A (en) * | 2021-06-30 | 2021-09-24 | 浙江大华技术股份有限公司 | Traffic marking detection method, device, equipment and medium |
CN112241714B (en) * | 2020-10-22 | 2024-04-26 | 北京字跳网络技术有限公司 | Method and device for identifying designated area in image, readable medium and electronic equipment |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104809433A (en) * | 2015-04-21 | 2015-07-29 | 电子科技大学 | Zebra stripe detection method based on maximum stable region and random sampling |
CN105718870A (en) * | 2016-01-15 | 2016-06-29 | 武汉光庭科技有限公司 | Road marking line extracting method based on forward camera head in automatic driving |
CN105893949A (en) * | 2016-03-29 | 2016-08-24 | 西南交通大学 | Lane line detection method under complex road condition scene |
JP2017223511A (en) * | 2016-06-14 | 2017-12-21 | 日本電信電話株式会社 | Road structuring device, road structuring method and road structuring program |
-
2019
- 2019-03-12 CN CN201910185277.9A patent/CN111695373B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104809433A (en) * | 2015-04-21 | 2015-07-29 | 电子科技大学 | Zebra stripe detection method based on maximum stable region and random sampling |
CN105718870A (en) * | 2016-01-15 | 2016-06-29 | 武汉光庭科技有限公司 | Road marking line extracting method based on forward camera head in automatic driving |
CN105893949A (en) * | 2016-03-29 | 2016-08-24 | 西南交通大学 | Lane line detection method under complex road condition scene |
JP2017223511A (en) * | 2016-06-14 | 2017-12-21 | 日本電信電話株式会社 | Road structuring device, road structuring method and road structuring program |
Non-Patent Citations (2)
Title |
---|
孙伟;张小瑞;唐慧强;张为公;闾军;: "Hough变换和最小二乘拟合的车道线协调检测" * |
闫利;黄亮;: "街景影像中的斑马线检测与定位" * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112241714A (en) * | 2020-10-22 | 2021-01-19 | 北京字跳网络技术有限公司 | Method and device for identifying designated area in image, readable medium and electronic equipment |
CN112241714B (en) * | 2020-10-22 | 2024-04-26 | 北京字跳网络技术有限公司 | Method and device for identifying designated area in image, readable medium and electronic equipment |
CN112446315A (en) * | 2020-11-19 | 2021-03-05 | 武汉中海庭数据技术有限公司 | High-precision map zebra crossing fusion method, system, server and storage medium |
CN112446315B (en) * | 2020-11-19 | 2022-05-31 | 武汉中海庭数据技术有限公司 | High-precision map zebra crossing fusion method, system, server and storage medium |
CN112633151A (en) * | 2020-12-22 | 2021-04-09 | 浙江大华技术股份有限公司 | Method, device, equipment and medium for determining zebra crossing in monitored image |
CN112633151B (en) * | 2020-12-22 | 2024-04-12 | 浙江大华技术股份有限公司 | Method, device, equipment and medium for determining zebra stripes in monitoring images |
CN113435350A (en) * | 2021-06-30 | 2021-09-24 | 浙江大华技术股份有限公司 | Traffic marking detection method, device, equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN111695373B (en) | 2024-02-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
US8340420B2 (en) | Method for recognizing objects in images | |
CN109978839B (en) | Method for detecting wafer low-texture defects | |
CN106683119B (en) | Moving vehicle detection method based on aerial video image | |
CN112686812B (en) | Bank card inclination correction detection method and device, readable storage medium and terminal | |
CN108305260B (en) | Method, device and equipment for detecting angular points in image | |
CN106709500B (en) | Image feature matching method | |
CN109413411B (en) | Black screen identification method and device of monitoring line and server | |
CN109685045A (en) | A kind of Moving Targets Based on Video Streams tracking and system | |
CN111695373B (en) | Zebra stripes positioning method, system, medium and equipment | |
CN113052170B (en) | Small target license plate recognition method under unconstrained scene | |
Xiang et al. | Moving object detection and shadow removing under changing illumination condition | |
CN111382658B (en) | Road traffic sign detection method in natural environment based on image gray gradient consistency | |
CN113435407B (en) | Small target identification method and device for power transmission system | |
Naufal et al. | Preprocessed mask RCNN for parking space detection in smart parking systems | |
CN108073940A (en) | A kind of method of 3D object instance object detections in unstructured moving grids | |
Ghahremannezhad et al. | Automatic road detection in traffic videos | |
CN111401364A (en) | License plate positioning algorithm based on combination of color features and template matching | |
CN111695374B (en) | Segmentation method, system, medium and device for zebra stripes in monitoring view angles | |
TWI498830B (en) | A method and system for license plate recognition under non-uniform illumination | |
Pratomo et al. | Parking detection system using background subtraction and HSV color segmentation | |
CN116740758A (en) | Bird image recognition method and system for preventing misjudgment | |
Kumar et al. | Intelligent parking vehicle identification and classification system | |
CN114927236A (en) | Detection method and system for multiple target images | |
CN110334703B (en) | Ship detection and identification method in day and night image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |