CN109284701A - A kind of driving recognition methods based on regional correlation - Google Patents
A kind of driving recognition methods based on regional correlation Download PDFInfo
- Publication number
- CN109284701A CN109284701A CN201811038594.XA CN201811038594A CN109284701A CN 109284701 A CN109284701 A CN 109284701A CN 201811038594 A CN201811038594 A CN 201811038594A CN 109284701 A CN109284701 A CN 109284701A
- Authority
- CN
- China
- Prior art keywords
- image
- driving
- recognition methods
- region
- methods based
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- 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
- G06V10/443—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 by matching or filtering
Abstract
The driving recognition methods based on regional correlation that the present invention relates to a kind of includes the following steps: that (1) acquires image;(2) acquired image is pre-processed;(3) according to the ambient conditions of driving conditions, sample characteristics data set is established, modeling training and study are carried out by data set, predicted with new input;(4) multiple regions are marked off according to HOG feature by pretreated image;(5) it is predicted according to data set, each region marked off is subjected to screening and micronization processes, depicts each area-of-interest profile;(6) area-of-interest is numbered, and detects the relative velocity in each region, divided the area into multiple set according to the difference of relative velocity size and sort out, and number;(7) ingredient of each set is analyzed, the big set of region quantity is classified as environment set, other set are classified as driving pedestrian's set.The present invention has the advantage that respective reaction is made convenient for identifying rows man-powered vehicle and rapidly, to reduce the risk of nighttime driving.
Description
Technical field
The present invention relates to the intelligent identifying system technical field of road travel, in particular to a kind of rows based on regional correlation
Vehicle recognition methods.
Background technique
With automobile in people's lives a large amount of universal, the risk of road travel is also being gradually increased, especially night
Driving.It is reported that the nocturnal traffic accident in China, wherein having 70% or so is to be made since driver is violating the regulations using light
Dazzle the eyes at light and can not see clearly caused by road surface ahead situation.Current vehicle exterior car lamp device substantially all relies on
In manual operation, in night running, the operation of multiple car lights is extremely cumbersome, and misoperation can inevitably occur or maloperation is existing
As.Therefore, being controlled using accurately and effectively intelligent automotive light has important social effect to safe driving.According to domestic and international at present
Present Research has begun on the high-grade car of benz BMW etc. using intelligent matrix headlamp, and principle is exactly to pass through
Multiple LED lamplights of design matrix arrangement are adjusted corresponding according to the orientation of other drivings or pedestrian for detecting in running car
LED light source, illumination to target bearing is closed, to reach abblendbar effect.Just because of this, identification is driven a vehicle and is made rapidly
The relevant technologies of respective reaction are particularly important out.
Summary of the invention
The technical problem to be solved by the present invention is to provide one kind for the concrete condition of the Road Detection in driving at night
The method of image recognition fully considers the complex situations of environmental factor, automatic identification image, to reduce the danger of nighttime driving
Property.
In order to solve the above technical problems, technical solution provided by the invention are as follows: a kind of driving identification based on regional correlation
Method includes the following steps:
(1) image is acquired, the road image in driving front is shot using CCD camera;
(2) image preprocessing is carried out to acquired image, concrete operations include image gray processing, grey level enhancement, filter
Wave processing, Morphological scale-space;
(3) it establishes recognition methods: according to the ambient conditions of driving conditions, certain sample characteristics number is established by testing
According to collection, the sample of data includes all kinds of drivings, road sign, roadside greening, street lamp, pedestrian etc., carries out modeling training by data set
And study, it can be predicted in practical applications with new input;
(4) multiple regions are marked off according to HOG feature by pretreated image;
(5) it is predicted according to data set, each region marked off is subjected to screening and micronization processes, it is emerging to depict each sense
Interesting region contour;
(6) area-of-interest is numbered, and detects the relative velocity in each region, then with algorithm to obtained number
According to error concealment is carried out, multiple set are divided the area into according to the difference of relative velocity size and are sorted out, and are numbered;
(7) ingredient of each set is analyzed, the big set of region quantity is classified as environment set, other set are classified as driving pedestrian
Set.The principle of analysis aggregate type is: will choose the stationary objects such as label, trees, street lamp to detect the region of relative velocity
It is attributed to environment set, dynamic object will be chosen to detect the region of relative velocity and be attributed to driving pedestrian's set, because quiet in environment
State object is more than dynamic object, therefore the region quantity in environment set will be significantly greater tnan the number of regions in driving pedestrian's set
Amount, to determine aggregate type according to region quantity.
As an improvement, the data-handling procedure of relative velocity uses the related theory of fuzzy algorithmic approach, set according to the actual situation
Set certain weight relationship, determine two data magnitude differences can be assumed that in a certain range be it is equal, to eliminate
The data fluctuations as caused by a variety of errors.
As an improvement, the mean value method that image gray processing uses, the i.e. three primary colors such as gray value after color image gray processing
The average value of channel luminance obtains gray level image.
As an improvement, the image after image gray processing carries out grey level enhancement, using binary conversion treatment, number two is obtained
Value image.
As an improvement, executing relevant operation in the filter process to image using Sigma filter value, being dropped
It makes an uproar filtering image.
The present invention has the advantage that respective reaction is made convenient for identifying rows man-powered vehicle and rapidly, to reduce nighttime driving
Risk.
Detailed description of the invention
Fig. 1 is a kind of structural block diagram of the driving recognition methods based on regional correlation of the present invention.
Specific embodiment
Present invention will be described in further detail below with reference to the accompanying drawings.
In conjunction with attached drawing, a kind of driving recognition methods based on regional correlation includes the following steps:
(1) image is acquired, the road image in driving front is shot using CCD camera;
(2) image preprocessing is carried out to acquired image, concrete operations include image gray processing, grey level enhancement, filter
Wave processing, Morphological scale-space;
(3) it establishes recognition methods: according to the ambient conditions of driving conditions, certain sample characteristics number is established by testing
According to collection, the sample of data includes all kinds of drivings, road sign, roadside greening, street lamp, pedestrian etc., carries out modeling training by data set
And study, it can be predicted in practical applications with new input;
(4) multiple regions are marked off according to HOG feature by pretreated image;
(5) it is predicted according to data set, each region marked off is subjected to screening and micronization processes, it is emerging to depict each sense
Interesting region contour;
(6) area-of-interest is numbered, and detects the relative velocity in each region, then with algorithm to obtained number
According to error concealment is carried out, multiple set are divided the area into according to the difference of relative velocity size and are sorted out, and are numbered;
(7) ingredient of each set is analyzed, the big set of region quantity is classified as environment set, other set are classified as driving pedestrian
Set.The principle of analysis aggregate type is: will choose the stationary objects such as label, trees, street lamp to detect the region of relative velocity
It is attributed to environment set, dynamic object will be chosen to detect the region of relative velocity and be attributed to driving pedestrian's set, because quiet in environment
State object is more than dynamic object, therefore the region quantity in environment set will be significantly greater tnan the number of regions in driving pedestrian's set
Amount, to determine aggregate type according to region quantity.
The data-handling procedure of relative velocity uses the related theory of fuzzy algorithmic approach, is acquired by mass data and according to reality
Certain weight relationship is arranged in border situation, can calculate the weight ratio of related data, therefore, it is determined that two data magnitude differences exist
It is a certain range of can be assumed that be it is equal, to eliminate the data fluctuations as caused by a variety of errors.
The mean value method that image gray processing uses, i.e. the primary display channels brightness such as gray value after color image gray processing
Average value obtains gray level image, and the image after image gray processing carries out grey level enhancement, using binary conversion treatment, in ash
Image extracting waste gray value after degreeization is 1, and black gray value is 0, two-value digitalization image is obtained, at the filtering to image
During reason, relevant operation is executed using Sigma filter value, obtains noise reduction filtering image.To morphological image process process
In, according to the theory of set theory in mathematical mor-phology, using corresponding in certain element structure form Detection and Extraction image
Shape information, with realize identification.
It is as follows to the utilization of fuzzy algorithmic approach correlation theory in data handling: according to circumstances to assume, it is assumed that at automobile
In the section of speed limit 60km/m, and the travel speed of automobile at a certain moment is 50km/h.The basic domain of relative velocity: according to reality
Drive a vehicle situation, it is assumed that the travel speed to always vehicle and vehicle in the same direction detected in section all 20km/h~60km/h it
Between, it detects that the approximate velocity of stationary object between 0~5km/h, then carries out calculating conversion with this vehicle speed data, takes opposite
The basic domain of speed is [- 30,120], is quantified as 6 grades { -1,0,1,2,3,4 }, and the quantizing factor of relative velocity is 5/
150=0.033.Deviation linguistic variable is 5, i.e. { NB (negative big), NS (is born small), ZE (zero), PS (just small), PB (honest) }.
According to deviation linguistic variable, realization divides the area into five grades.Analyze the region quantity in five grades, quantity is most
Region in grade belongs to environment set, and the region in remaining grade belongs to driving pedestrian's set.
The present invention and its embodiments have been described above, this description is no restricted, shown in the drawings
Only one of embodiments of the present invention, actual structure is not limited to this.All in all if the ordinary skill of this field
Personnel are enlightened by it, without departing from the spirit of the invention, are not inventively designed and the technical solution phase
As frame mode and embodiment, be within the scope of protection of the invention.
Claims (5)
1. a kind of driving recognition methods based on regional correlation, characterized by the following steps:
(1) image is acquired, the road image in driving front is shot using CCD camera;
(2) image preprocessing is carried out to acquired image, concrete operations include image gray processing, grey level enhancement, at filtering
Reason, Morphological scale-space;
(3) it establishes recognition methods: according to the ambient conditions of driving conditions, certain sample characteristics data is established by testing
Collection, the samples of data include all kinds of drivings, road sign, roadside greening, street lamp, pedestrian etc., by data set carry out modeling training with
Study, can be predicted in practical applications with new input;
(4) multiple regions are marked off according to HOG feature by pretreated image;
(5) it is predicted according to data set, each region marked off is subjected to screening and micronization processes, depicts each region of interest
Domain profile;
(6) area-of-interest is numbered, and detects the relative velocity in each region, then with algorithm to obtained data into
Row error concealment divides the area into multiple set according to the difference of relative velocity size and sorts out, and numbers;
(7) ingredient of each set is analyzed, the big set of region quantity is classified as environment set, other set are classified as driving pedestrian's collection
It closes.
2. a kind of driving recognition methods based on regional correlation according to claim 1, it is characterised in that: relative velocity
Data-handling procedure uses the related theory of fuzzy algorithmic approach, and certain weight relationship is arranged according to the actual situation, determines two numbers
Be according to magnitude difference can be assumed that in a certain range it is equal, to eliminate the data fluctuations as caused by a variety of errors.
3. a kind of driving recognition methods based on regional correlation according to claim 1, it is characterised in that: image gray processing
The mean value method of use, the i.e. average value of the primary display channels brightness such as gray value after color image gray processing obtain gray processing
Image.
4. a kind of driving recognition methods based on regional correlation according to claim 1, it is characterised in that: by image ash
Image after degreeization carries out grey level enhancement, using binary conversion treatment, obtains two-value digitalization image.
5. a kind of driving recognition methods based on regional correlation according to claim 1, it is characterised in that: to image
In filter process, relevant operation is executed using Sigma filter value, obtains noise reduction filtering image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811038594.XA CN109284701A (en) | 2018-09-06 | 2018-09-06 | A kind of driving recognition methods based on regional correlation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811038594.XA CN109284701A (en) | 2018-09-06 | 2018-09-06 | A kind of driving recognition methods based on regional correlation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109284701A true CN109284701A (en) | 2019-01-29 |
Family
ID=65184154
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811038594.XA Pending CN109284701A (en) | 2018-09-06 | 2018-09-06 | A kind of driving recognition methods based on regional correlation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109284701A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110826445A (en) * | 2019-10-28 | 2020-02-21 | 衢州学院 | Method and device for detecting specific target area in colorless scene video |
CN111182693A (en) * | 2019-12-28 | 2020-05-19 | 杭州拓深科技有限公司 | Linkage induction type street lamp brightness adjusting system and control method thereof |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102592114A (en) * | 2011-12-26 | 2012-07-18 | 河南工业大学 | Method for extracting and recognizing lane line features of complex road conditions |
US20150139551A1 (en) * | 2013-11-15 | 2015-05-21 | Adobe Systems Incorporated | Cascaded Object Detection |
US20150154454A1 (en) * | 2013-05-16 | 2015-06-04 | Microsoft Technology Licensing, Llc | Motion stabilization and detection of articulated objects |
CN104751483A (en) * | 2015-03-05 | 2015-07-01 | 北京农业信息技术研究中心 | Method for monitoring abnormal conditions of working region of warehouse logistics robot |
US20160019427A1 (en) * | 2013-03-11 | 2016-01-21 | Michael Scott Martin | Video surveillence system for detecting firearms |
CN106022231A (en) * | 2016-05-11 | 2016-10-12 | 浙江理工大学 | Multi-feature-fusion-based technical method for rapid detection of pedestrian |
CN106114511A (en) * | 2016-07-21 | 2016-11-16 | 辽宁工业大学 | A kind of automobile cruise system core target identification method |
CN106530310A (en) * | 2016-10-25 | 2017-03-22 | 深圳大学 | Pedestrian counting method and device based on human head top recognition |
CN107316002A (en) * | 2017-06-02 | 2017-11-03 | 武汉理工大学 | A kind of night front vehicles recognition methods based on Active Learning |
CN108039046A (en) * | 2018-01-05 | 2018-05-15 | 重庆邮电大学 | A kind of city intersection pedestrian detection identifying system based on C-V2X |
CN108226924A (en) * | 2018-01-11 | 2018-06-29 | 李烜 | Running car environment detection method, apparatus and its application based on millimetre-wave radar |
CN108470460A (en) * | 2018-04-11 | 2018-08-31 | 江苏大学 | A kind of nearby vehicle Activity recognition method based on smart mobile phone and RNN |
-
2018
- 2018-09-06 CN CN201811038594.XA patent/CN109284701A/en active Pending
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102592114A (en) * | 2011-12-26 | 2012-07-18 | 河南工业大学 | Method for extracting and recognizing lane line features of complex road conditions |
US20160019427A1 (en) * | 2013-03-11 | 2016-01-21 | Michael Scott Martin | Video surveillence system for detecting firearms |
US20150154454A1 (en) * | 2013-05-16 | 2015-06-04 | Microsoft Technology Licensing, Llc | Motion stabilization and detection of articulated objects |
US20150139551A1 (en) * | 2013-11-15 | 2015-05-21 | Adobe Systems Incorporated | Cascaded Object Detection |
CN104751483A (en) * | 2015-03-05 | 2015-07-01 | 北京农业信息技术研究中心 | Method for monitoring abnormal conditions of working region of warehouse logistics robot |
CN106022231A (en) * | 2016-05-11 | 2016-10-12 | 浙江理工大学 | Multi-feature-fusion-based technical method for rapid detection of pedestrian |
CN106114511A (en) * | 2016-07-21 | 2016-11-16 | 辽宁工业大学 | A kind of automobile cruise system core target identification method |
CN106530310A (en) * | 2016-10-25 | 2017-03-22 | 深圳大学 | Pedestrian counting method and device based on human head top recognition |
CN107316002A (en) * | 2017-06-02 | 2017-11-03 | 武汉理工大学 | A kind of night front vehicles recognition methods based on Active Learning |
CN108039046A (en) * | 2018-01-05 | 2018-05-15 | 重庆邮电大学 | A kind of city intersection pedestrian detection identifying system based on C-V2X |
CN108226924A (en) * | 2018-01-11 | 2018-06-29 | 李烜 | Running car environment detection method, apparatus and its application based on millimetre-wave radar |
CN108470460A (en) * | 2018-04-11 | 2018-08-31 | 江苏大学 | A kind of nearby vehicle Activity recognition method based on smart mobile phone and RNN |
Non-Patent Citations (5)
Title |
---|
LIU BO 等: "USING OBJECT CLASSIFICATION TO IMPROVE URBAN TRAFFIC MONITORING SYSTEM", 《IEEE INT. CONF. NEURAL NETWORKS&SIGNAL PROCESSING》 * |
刘禹希: "基于机器视觉的智能驾驶车辆的目标识别研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
张晴 等: "基于区域对比的图像显著目标检测", 《上海应用技术学院学报(自然科学版)》 * |
王欢 等: "一种区域级运动目标检测方法", 《模式识别与人工智能》 * |
金春霞: "模糊逻辑理论及其在智能交通系统中的应用研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110826445A (en) * | 2019-10-28 | 2020-02-21 | 衢州学院 | Method and device for detecting specific target area in colorless scene video |
CN110826445B (en) * | 2019-10-28 | 2021-04-23 | 衢州学院 | Method and device for detecting specific target area in colorless scene video |
CN111182693A (en) * | 2019-12-28 | 2020-05-19 | 杭州拓深科技有限公司 | Linkage induction type street lamp brightness adjusting system and control method thereof |
CN111182693B (en) * | 2019-12-28 | 2022-04-15 | 杭州拓深科技有限公司 | Linkage induction type street lamp brightness adjusting system and control method thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107766821B (en) | Method and system for detecting and tracking full-time vehicle in video based on Kalman filtering and deep learning | |
CN102637257B (en) | Video-based detection and recognition system and method of vehicles | |
CN102968646B (en) | A kind of detection method of license plate based on machine learning | |
CN101783076B (en) | Method for quick vehicle type recognition under video monitoring mode | |
CN100495438C (en) | Method for detecting and identifying moving target based on video monitoring | |
CN105913041A (en) | Pre-marked signal lights based identification method | |
CN202084185U (en) | Automatic traffic sign identification device | |
Suhweil et al. | Smart controlling for traffic light time | |
CN103034843B (en) | Method for detecting vehicle at night based on monocular vision | |
CN104992145A (en) | Moment sampling lane tracking detection method | |
CN103400111A (en) | Method for detecting fire accident on expressway or in tunnel based on video detection technology | |
CN103366190A (en) | Method for identifying traffic sign | |
CN104408931A (en) | Incomplete sign license plate identification system and method | |
CN103279756A (en) | Vehicle detecting analysis system and detecting analysis method thereof based on integrated classifier | |
CN1932838A (en) | Vehicle plate extracting method based on skiagraphy and mathematical morphology | |
CN102298693A (en) | Expressway bend detection method based on computer vision | |
CN102110366B (en) | Block-based accumulated expressway vehicle parking event detecting method | |
CN105760847A (en) | Visual detection method for detecting whether motor cyclist wears helmet or not | |
CN109284701A (en) | A kind of driving recognition methods based on regional correlation | |
CN103927548A (en) | Novel vehicle collision avoiding brake behavior detection method | |
Xing et al. | Traffic sign detection and recognition using color standardization and Zernike moments | |
CN104992160B (en) | A kind of heavy truck night front vehicles detection method | |
CN107545244A (en) | Speed(-)limit sign detection method based on image processing techniques | |
Wen et al. | A rear-vehicle detection system for static images based on monocular vision | |
CN105046223A (en) | Device for detecting severity of ''black-hole effect'' at tunnel entrance and method thereof |
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 | ||
AD01 | Patent right deemed abandoned | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20220614 |