CN103077540A - Automatic detection method of vehicle moving region in video monitoring of express way - Google Patents
Automatic detection method of vehicle moving region in video monitoring of express way Download PDFInfo
- Publication number
- CN103077540A CN103077540A CN2013100367270A CN201310036727A CN103077540A CN 103077540 A CN103077540 A CN 103077540A CN 2013100367270 A CN2013100367270 A CN 2013100367270A CN 201310036727 A CN201310036727 A CN 201310036727A CN 103077540 A CN103077540 A CN 103077540A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- space
- time
- obtains
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 22
- 238000001514 detection method Methods 0.000 title abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 13
- 230000007797 corrosion Effects 0.000 claims abstract description 5
- 238000005260 corrosion Methods 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 24
- 238000006243 chemical reaction Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 239000012634 fragment Substances 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 5
- 238000011160 research Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an automatic detection method of a vehicle moving region in the video monitoring of an express way. The automatic detection method is used for automatically detecting a vehicle moving region in a video on the basis of a frequency domain analysis theory and comprises the following steps of: constructing a three-dimensional vector space in terms of time and space; carrying out Fourier transform on time domain information of each pixel; acquiring a maximum mono-frequency power; carrying out binaryzation, corrosion and expansion processing on a transformed image; determining the boundary of the binarized image; and finally acquiring the vehicle moving region by using the maximum mono-frequency power value. The automatic detection method disclosed by the invention has the advantages of easy implementation, strong robustness, high accuracy and the like and can be used for providing effective auxiliary information for vehicle information and moving conditions in the video monitoring process.
Description
Technical field
The invention belongs to technical field of video monitoring, vehicle movement zone automatic testing method in the video monitoring of especially a kind of through street, the method utilizes frequency domain analysis theoretical, and a video middle rolling car moving region is detected automatically.
Background technology
Along with the development of Video Supervision Technique, video camera has been widely used for carries out Real Time Monitoring to various environment, zone and place.Along with the rapid increase of video detector quantity, traditional artificial passive monitoring can't satisfy the needs of monitor task far away.Therefore, realization can replace the intelligent automatic monitoring function of human eye to become the target of video monitoring research.At present, in the research of Intelligent traffic video supervisory system, be the hot issue of research to automatic detection, the recognition and tracking of driving vehicle on the highway always, and also produced gradually relevant monitoring product on the market.Yet existing monitoring product mostly is for concrete monitoring position, and video camera is carried out ad hoc Installation and Debugging, and recycles the video that collects afterwards, and video camera is demarcated processing further.This use procedure greatly reduces the efficient of installing and using of rig camera, and after video camera damages replacing or shooting angle variation, also needs again it to be demarcated, and has also increased the cost of safeguarding.Therefore, this needs manually carry out the situation of prior calibrating camera, are one of major obstacles that finally realizes vehicle intelligent video monitoring system widespread use on the highway.
Morris and Trivedi propose to adopt the method for vehicle tracking track to come automatic acquisition Vehicle Driving Cycle zone in " Learning; modeling; and classification of vehicle track patterns from live video (study of vehicle tracking pattern, modeling and classification in the live video); IEEE Transactions on Intelligent Transportation Systems (IEEE intelligent transportation system proceedings), 2008 ".Yet, this method is in time domain picture signal to be processed, the method easily is subject to the interference of noise in time domain signal, the number of following the tracks of vehicle is had relatively high expectations, and also be difficult to obtain accurate, complete vehicle movement zone by track of vehicle, the accuracy of follow-up vehicle identification and event detection is caused bad impact.
Summary of the invention
The object of the invention is to overcome deficiency of the prior art, vehicle movement zone automatic testing method in the video monitoring of through street is provided in a kind of video, can be under the complex environment of the practical engineering application such as illumination variation, detection noise, obtain rapidly and accurately the zone that vehicle movement passes through, i.e. highway region in the video image.
For achieving the above object, the video that successive frame is at first read in the present invention forms the time of pixel and the tri-vector space in space, then time-varying information corresponding to each pixel done Fourier transform, obtain the power spectrum of frequency-region signal from frequency domain information, generate changing image according to its unifrequency peak power on certain frequency, then this image is carried out binaryzation and corrosion expansive working, delimit at last the road area of rectangle.
Vehicle movement zone automatic testing method may further comprise the steps in a kind of through street video monitoring that the present invention proposes:
Step 1, each frame to the monitor video fragment carries out the Gaussian Blur processing, obtain comprising accordingly with this frame the tri-vector space in time and space, and described tri-vector space is stored as two-dimension time-space t-y image by row, wherein, t represents the time take frame as unit, is the horizontal ordinate of described space-time t-y image, y represents the ordinate of pixel in the row that extract, and is the ordinate of described space-time t-y image;
Step 2 is carried out Fourier transform to the time-domain information of each pixel in the described two-dimension time-space t-y image, obtains corresponding frequency domain information F (u);
Step 3 obtains the maximum unifrequency power at each pixel place according to resulting frequency domain information F (u), according to the correspondence position of pixel, obtain the image behind the described two-dimension time-space t-y image conversion;
Step 4 is carried out binaryzation, corrosion and expansion process successively to the image after the described conversion;
Step 5 is determined the border through the image that obtains after described step 4 processing, obtains the vehicle movement zone in described each frame of monitor video fragment.
The present invention's remarkable result compared with prior art is: the moving region of vehicle in can the automatic acquisition video monitoring, and the interference such as the illumination variation that occurs in the actual monitored video, DE Camera Shake, detection noise, foreground detection error are had higher robustness simultaneously.Owing to having overcome these prior aries insoluble difficulty aspect practical engineering application, therefore really having realized the fast detecting in vehicle movement zone in the video.
The present invention is directed to the needs that vehicle target is identified and followed the tracks of in the highway communication intelligent video monitoring, utilize the frequency-domain analysis technology, detect the moving region of vehicle, have the advantages such as algorithm is simple, degree of accuracy is high, strong robustness.The present invention uses the image processing techniques based on frequency-domain analysis, realizes the automatic detection in vehicle movement zone, finally provides a convenient and reliable highway zone location algorithm for vehicle target identification and tracking in the highway communication intelligent video monitoring.
Description of drawings
Fig. 1 is vehicle movement zone automatic testing method process flow diagram in the video monitoring of through street of the present invention;
Fig. 2 is the sectional drawing of used video according to an embodiment of the invention;
Fig. 3 has shown the t-y image of a certain row in the video according to an embodiment of the invention;
Fig. 4 obtains power spectrum chart for carrying out Fourier transform by the t-y image to Fig. 3;
Fig. 5 is for depositing back the unifrequency peak power image after the conversion of former pixel correspondence position;
Fig. 6 is that design sketch is detected in final vehicle movement zone.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.
Fig. 1 is vehicle movement zone automatic testing method process flow diagram in the video monitoring of through street of the present invention.The video clips that the present invention chooses one section rig camera shooting is that the present invention will be described for example, enough vehicles are arranged from each track process in this video, and break in traffic rules and regulations and accident do not occur, Fig. 2 is the sectional drawing of used video in this embodiment of the invention, and reads in continuously the appointment frame number of described video in the gray scale mode.As shown in Figure 1, the method may further comprise the steps:
Step 1, each frame to the monitor video fragment carries out the Gaussian Blur processing, obtain comprising accordingly with this frame the tri-vector space in time and space, and described tri-vector space is stored as two-dimension time-space t-y image by row, wherein, t represents the time take frame as unit, horizontal ordinate for described image, y represents the ordinate of pixel in the row that extract, and is the ordinate of described image, and Fig. 3 has shown the wherein t-y element of row of this image;
Described Gaussian Blur is treated to the use Gaussian function
Sampling value each pixel is carried out convolution operation, wherein, x, y are the side-play amount of the relative center pixel of pixel, σ is the standard deviation of these side-play amounts.If position (x, y) locates the pixel value of the pixel of t frame is A (x, y, t), the value of pixel in spacetime coordinates after then processing through above-mentioned Gaussian Blur is:
Step 2 is carried out Fourier transform to the time-domain information of each pixel in the described two-dimension time-space t-y image, obtains corresponding frequency domain information;
In this step, pixel (x
0, y
0) time series on time shaft is f (t)=B (x
0, y
0, t), this sequence is done Fourier transform, can obtain the expression of this time series f (t) under frequency domain, namely with the corresponding frequency-region signal F of described time series (u):
Wherein, N is sampling sum, herein total number of image frames of corresponding intercepting;
Step 3 obtains the maximum unifrequency power at each pixel place according to resulting frequency domain information, according to the correspondence position of pixel, obtain the image after the conversion;
This step is further comprising the steps:
Step 31 obtains the relative power P (u) of its each frequency component from the value of the frequency-region signal F (u) of each pixel:
P(u)＝|F(u)|
2；
Fig. 4 for ease of showing, has carried out a nonlinear transformation for the result after carrying out Fourier transform based on the shown sectional drawing of Fig. 2 and asking for the relative power of each frequency component in Fig. 4.
Step 32 travels through described relative power P (u), selects maximum relative power m from the relative power that is higher than assigned frequency:
m＝max{P(u)|u＞u
thr}， u＝0，1，2，3，…，N-1；
Wherein, the relative power that P (u) draws for previous step, u
ThrLow-limit frequency for appointment.
In an embodiment of the present invention, can begin to travel through P (u) from u 〉=0.02N.
Step 33, each pixel in the described two-dimension time-space t-y image is carried out respectively the processing of described step 31 and 32, and maximal phase corresponding to each pixel that will obtain deposits correspondence position in the image in to performance number, obtain the image after the conversion, i.e. each in the image after this conversion point M (x, y) be the m value that the point (x, y) in the described two-dimension time-space t-y image produces in described step 32, as shown in Figure 5.
Step 4, the image after the described conversion is carried out binaryzation, corrosion and expansion process successively:
In this step, the every bit that travels through in the image after the described conversion carries out binary conversion treatment: all numerical value is made as white greater than the pixel of assign thresholds, and namely 1, other point is made as black, and namely 0.Then the bianry image that generates is successively once corroded and expansion process.
Step 5 is determined the border through the image that obtains after described step 4 processing, obtains the vehicle movement zone in described each frame of monitor video fragment.
In this step, begin inwardly to do the straight line that is parallel to four limits from four limits of passing through the image that obtains after described step 4 is processed, make the center of the constantly close image of these straight lines, until obtain to surround the minimum rectangle of all non-zero pixels, so just formed the square frame on a border, comprised the scope of through street in this square frame, wherein all non-zero points are exactly the moving region of vehicle in this image, the result who finally obtains as shown in Figure 6, the square frame among Fig. 6 is the moving region of vehicle in the image.
Above-described specific embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (9)
1. vehicle movement zone automatic testing method in the through street video monitoring is characterized in that, the method may further comprise the steps:
Step 1, each frame to the monitor video fragment carries out the Gaussian Blur processing, obtain comprising accordingly with this frame the tri-vector space in time and space, and described tri-vector space is stored as two-dimension time-space t-y image by row, wherein, t represents the time take frame as unit, is the horizontal ordinate of described space-time t-y image, y represents the ordinate of pixel in the row that extract, and is the ordinate of described space-time t-y image;
Step 2 is carried out Fourier transform to the time-domain information of each pixel in the described two-dimension time-space t-y image, obtains corresponding frequency domain information F (u);
Step 3 obtains the maximum unifrequency power at each pixel place according to resulting frequency domain information F (u), according to the correspondence position of pixel, obtain the image behind the described two-dimension time-space t-y image conversion;
Step 4 is carried out binaryzation, corrosion and expansion process successively to the image after the described conversion;
Step 5 is determined the border through the image that obtains after described step 4 processing, obtains the vehicle movement zone in described each frame of monitor video fragment.
2. method according to claim 1 is characterized in that, described Gaussian Blur is treated to the use Gaussian function
Sampling value each pixel is carried out convolution operation, wherein, x, y are the side-play amount of the relative center pixel of pixel, σ is the standard deviation of these side-play amounts.
3. method according to claim 2 is characterized in that, is A (x, y, t) if position (x, y) locates the pixel value of the pixel of t frame, and the value of pixel in spacetime coordinates after then processing through described Gaussian Blur is:
4. method according to claim 1 is characterized in that, in the described step 2, if pixel (x
0, y
0) time series on time shaft is f (t)=B (x
0, y
0, t), then described time series is carried out Fourier transform, obtain the expression F (u) of described time series under frequency domain:
Wherein, N is sampling sum, i.e. total number of image frames.
5. method according to claim 1 is characterized in that, described step 3 is further comprising the steps:
Step 31, the relative power P (u) of each frequency component of acquisition from the frequency domain information of a pixel:
P(u)＝|F(u)|
2；
Step 32 travels through described relative power P (u), selects maximum relative power m from the relative power that is higher than assigned frequency:
m＝max{P(u)|u＞u
thr}， u＝0，1，2，3，…，N-1；
Wherein, u
ThrBe the low-limit frequency of appointment, N is sampling sum, i.e. total number of image frames;
Step 33, each pixel in the described two-dimension time-space t-y image is carried out respectively the processing of described step 31 and 32, and maximal phase corresponding to each pixel that will obtain deposit correspondence position in the image in to performance number, obtains the image behind the described two-dimension time-space t-y image conversion.
6. method according to claim 5 is characterized in that, in the described step 32, begins to travel through described relative power P (u) from u 〉=0.02N.
7. method according to claim 5 is characterized in that, each the some M (x, y) in the image after the described conversion is the maximum relative power m value of the point (x, y) in the described two-dimension time-space t-y image.
8. method according to claim 1 is characterized in that, the binary conversion treatment in the described step 4 is: all numerical value is made as white greater than the pixel of assign thresholds, and namely 1, other point is made as black, and namely 0.
9. method according to claim 1, it is characterized in that, in the described step 5, begin inwardly to do the straight line that is parallel to four limits from four limits of passing through the image that obtains after described step 4 is processed, make the center of the constantly close image of these straight lines, until obtain surrounding the minimum rectangle of all non-zero pixels, form the square frame on a border, all non-zero points are the moving region of vehicle in this image in this square frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100367270A CN103077540A (en) | 2013-01-30 | 2013-01-30 | Automatic detection method of vehicle moving region in video monitoring of express way |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100367270A CN103077540A (en) | 2013-01-30 | 2013-01-30 | Automatic detection method of vehicle moving region in video monitoring of express way |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103077540A true CN103077540A (en) | 2013-05-01 |
Family
ID=48154059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013100367270A Pending CN103077540A (en) | 2013-01-30 | 2013-01-30 | Automatic detection method of vehicle moving region in video monitoring of express way |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103077540A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105303540A (en) * | 2015-08-14 | 2016-02-03 | 深圳市瀚海基因生物科技有限公司 | Single-molecule image correction device |
| CN112465850A (en) * | 2020-12-08 | 2021-03-09 | 中国科学院计算技术研究所数字经济产业研究院 | Peripheral boundary modeling method, intelligent monitoring method and device |
-
2013
- 2013-01-30 CN CN2013100367270A patent/CN103077540A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 孔庆杰: "信息融合理论及其在交通监控信息处理中的应用", 《CNKI中国博士学位论文全文库》, no. 10, 15 October 2010 (2010-10-15), pages 113 - 114 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105303540A (en) * | 2015-08-14 | 2016-02-03 | 深圳市瀚海基因生物科技有限公司 | Single-molecule image correction device |
| CN112465850A (en) * | 2020-12-08 | 2021-03-09 | 中国科学院计算技术研究所数字经济产业研究院 | Peripheral boundary modeling method, intelligent monitoring method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shim et al. | Road damage detection using super-resolution and semi-supervised learning with generative adversarial network | |
| US11783452B2 (en) | Traffic monitoring using distributed fiber optic sensing | |
| Li et al. | Road lane detection with gabor filters | |
| CN106127113A (en) | A kind of road track line detecting method based on three-dimensional laser radar | |
| CN106128121B (en) | Vehicle queue length fast algorithm of detecting based on Local Features Analysis | |
| CN103176185A (en) | Method and system for detecting road barrier | |
| CN104063882B (en) | Vehicle video speed measuring method based on binocular camera | |
| CN102592454A (en) | Intersection vehicle movement parameter measuring method based on detection of vehicle side face and road intersection line | |
| CN103077387B (en) | Carriage of freight train automatic testing method in video | |
| CN104050477A (en) | Infrared image vehicle detection method based on auxiliary road information and significance detection | |
| CN107516423B (en) | Video-based vehicle driving direction detection method | |
| CN109446917A (en) | A kind of vanishing Point Detection Method method based on cascade Hough transform | |
| CN103049788B (en) | Based on space number for the treatment of object detection system and the method for computer vision | |
| CN105321189A (en) | Complex environment target tracking method based on continuous adaptive mean shift multi-feature fusion | |
| CN112261719B (en) | Area positioning method combining SLAM technology with deep learning | |
| CN104899895A (en) | Detection method of trace complexity of mobile targets of fire video in channel of power transmission line | |
| CN107944354A (en) | A kind of vehicle checking method based on deep learning | |
| Qi et al. | Small infrared target detection utilizing local region similarity difference map | |
| CN118096815A (en) | Road abnormal event detection system based on machine vision | |
| Qu et al. | Scale self-adaption tracking method of Defog-PSA-Kcf defogging and dimensionality reduction of foreign matter intrusion along railway lines | |
| Wang et al. | 3d reconstruction and segmentation system for pavement potholes based on improved structure-from-motion (sfm) and deep learning | |
| CN103077540A (en) | Automatic detection method of vehicle moving region in video monitoring of express way | |
| CN103310465A (en) | Vehicle occlusion treating method based on Markov random field | |
| Yu et al. | Adaptive vehicle extraction in real-time traffic video monitoring based on the fusion of multi-objective particle swarm optimization algorithm | |
| CN104240268A (en) | Pedestrian tracking method based on manifold learning and sparse representation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130501 |