CN102073846B - Method for acquiring traffic information based on aerial images - Google Patents
Method for acquiring traffic information based on aerial images Download PDFInfo
- Publication number
- CN102073846B CN102073846B CN2010105888800A CN201010588880A CN102073846B CN 102073846 B CN102073846 B CN 102073846B CN 2010105888800 A CN2010105888800 A CN 2010105888800A CN 201010588880 A CN201010588880 A CN 201010588880A CN 102073846 B CN102073846 B CN 102073846B
- Authority
- CN
- China
- Prior art keywords
- aircraft
- vehicle
- patch
- road
- aerial images
- 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.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
- G08G1/012—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from other sources than vehicle or roadside beacons, e.g. mobile networks
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0133—Traffic data processing for classifying traffic situation
Landscapes
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method for acquiring traffic information based on aerial images, which is characterized in that dynamic targets and static targets are detected in accordance with the analysis of the aerial images, wherein, in a method for detecting the dynamic targets, a Kanade-Lucas-Tomasi (KLT) algorithm is used to acquire a plurality of characteristic points and the motion parameters of the dynamic targets; motion direction vectors are determined by projection of road region directions; and a k-medoids algorithm is adopted to cluster the plurality of characteristic points, thus the dynamic targets are separated by the characteristic points. In a method for detecting the static targets, a road region restriction and plaque analysis method is used to define road regions and the static targets. On the basis of acquiring the traffic information, the information is comprehensively analyzed and processed, so that the comprehensive traffic state and the detailed traffic parameter indexes can be obtained.
Description
Technical field
The invention belongs to technical field of image processing, relate to a kind of method for acquiring traffic information based on Aerial Images and system.
Background technology
It is the basis of road grid traffic monitoring and early warning that sparse road net traffic state detects.For the traffic hazard of traffic events and other etc. is carried out science, evaluation and test and early warning rapidly, at first to carry out function upgrading to Traffic Video Detection System equipment, enable to adapt to Special Geographic, the natural conditions of western sparse road network.Then the traffic behavior that needs to study based on video detects and recognizer, realizes effective detection of traffic events, thereby provides reliable information for the road network safe early warning.
Because sparse road net traffic state is more special, the vehicle flowrate less does not have again enough personnel to go on-site supervision, can monitor this regional system and solves following problem so need badly:
(1) need to monitor its traffic conditions for the road area of no worker monitor, reduce the loss that traffic faults brings as far as possible.
(2) by Based Intelligent Control, reduce human resources and take, to optimize the manpower configuration.
Summary of the invention
Purpose of the present invention provides a kind of method for acquiring traffic information based on Aerial Images, overcomes inapplicable for dynamic background of traditional background subtraction based on plaque detection and tracking, the place of applicable change of background and light traffic.
For reaching above purpose, solution of the present invention is:
A kind of method for acquiring traffic information based on Aerial Images is analyzed according to Aerial Images, detects dynamic object and static object, extracts road traffic parameter;
Wherein, the dynamic object detection method is for adopting the KLT algorithm to obtain the kinematic parameter of some unique points and dynamic object, determine the direction of motion vector by the road area direction projection, adopt the k-medoids algorithm with some feature points clusterings, thereby by unique point, dynamic object is separated;
Static object detection method is for adopting road area restriction and patch analytical approach to limit road area and static object.
Further, described KLT algorithm is under the similar condition of the characteristic area gray scale between the Aerial Images consecutive frame, choose a large amount of unique points in grayscale image sequence, obtain characteristic point position thereby carry out the two dimensional character tracking, solve accordingly the method for two dimensional character kinematic parameter; Simultaneously, the projecting direction according to motion vector extracts road direction, and utilizes Hough transformation to detect Road.
Described k-medoids algorithm comprises the following steps:
(1) choose arbitrarily K object as medoids (O from unique point
1, O
2... O
i... O
k);
(2) object of remainder is assigned in each class according to the principle the most close with medoid gone;
(3) for each class (O
i) in, order is chosen an O
r, calculate and use O
rReplace O
iAfter consumption E (O
r), that O of selection E minimum
rReplace O
i
(4) return to step (2) cycle calculations, until K medoids is fixed up.
Described road area restriction is to adopt the mode of Color histogram distribution with the road area identification process, changes detected Road in conjunction with Hough and jointly limits road area.
Described patch analysis is the information such as the RGB information of the center point coordinate that specifically obtains patch, plaque area, patch and patch number, the position of the center point coordinate reflection static object of patch; Plaque area can be used for removing noise; The RGB information of patch is after the image that obtains of taking photo by plane carries out binary conversion treatment (namely being converted to the image that only has black and white two looks by coloured image), at this image) in determine the position of patch after, find corresponding patch position in the former Aerial Images, statistics RGB information is to distinguish target object and surrounding environment; The patch number is used for unified processing of patch to whole Aerial Images.
Described extraction road traffic parameter comprises extracts vehicle heading, car speed and length of wagon.
Described vehicle heading, to determine the travel direction of vehicle according to the magnitude relationship of vehicle movement vector and background dot motion vector, described vehicle movement vector refers to that dynamic object is with respect to the motion vector of aircraft, described background dot motion vector refers to static object with respect to the motion vector of aircraft, if the vehicle movement vector is greater than the background dot motion vector; Vehicle and aircraft reverse driving so; If the vehicle movement vector is less than the background dot motion vector, vehicle and aircraft travel in the same way so.
Described car speed calculates according to following formula,
When vehicle and aircraft reverse driving, V
Car=V
Relatively-V
Aircraft
When vehicle and aircraft travel in the same way, V
Car=V
Relatively+ V
Aircraft
Wherein, V
CarRefer to the speed of dynamic object, V
RelativelyRefer to the speed of static object, V
AircraftRefer to the speed of aircraft.
Described length of wagon is according to length=(y
max-y
min) * cos θ * (V
Plane* 0.04/vector
Background) calculating, wherein y
max-y
minBe the displacement of vehicle on the aircraft flight direction, θ is the angle of aircraft flight direction and vehicle heading, V
PlaneThe flying speed of aircraft, vector
BackgroundBe the speed of related movement of background dot on image, 0.04 is interval time between every frame.According to the characteristics of road target, respectively moving vehicle and stationary vehicle are detected.
The dynamic object detection method
For moving vehicle, adopt the KLT algorithm to detect, the KLT algorithm is under the similar condition of the characteristic area gray scale between the image consecutive frame, choose a large amount of unique points in grayscale image sequence, obtain characteristic point position thereby carry out the two dimensional character tracking, solve accordingly the method for two dimensional character kinematic parameter.
Simultaneously, according to the projecting direction of motion vector, can extract road direction, the error of bringing to reduce the aircraft shake.Utilize Hough transformation to detect Road.Is Hough transformation can detect all straight lines in image, and (method of Hough transformation detection of straight lines known in industry? preferably can simply introduce), calculate thus slope k and the intercept b of all straight lines, obtain the slope range of the maximum that distributes and obtain mean value k according to slope
avg, obtain thus direction and the screening of Road and fall other irrelevant straight lines.Obtain maximum intercept b by the gained intercept
maxIntercept b with minimum
minTwo borders of Road can be drawn, the target area can be roughly estimated thus.
In clustering algorithm, the most frequently used is the k-means algorithm.The k-means algorithm is accepted input quantity k; Then n data object is divided into k cluster in order to make the cluster that obtains satisfy: the object similarity in same cluster is higher; And the object similarity in different clusters is less.The cluster similarity is to utilize the average of object in each cluster to obtain " center object " (center of attraction) to calculate.
The course of work of k-means algorithm is described as follows: at first select arbitrarily k object as initial cluster center from n data object; And for other object of be left, according to the similarity (distance) of they and these cluster centres, respectively they are distributed to (cluster centre representative) cluster the most similar to it; And then calculate the cluster centre (average of all objects in this cluster) of each new cluster that obtains; Constantly repeat this process until the canonical measure function begins convergence.Generally all adopt mean square deviation to have following characteristics as .k cluster of canonical measure function: each cluster itself is compact as much as possible, and separates as much as possible between each cluster.
Can obtain the kinematic parameter of some unique points and target by the KLT algorithm, then determine the direction of motion vector by the road area direction projection, and the k-means algorithm be with some feature points clusterings, thereby by unique point, moving target is separated.
Static object detection method
For stationary vehicle, adopt the road area restriction to add the method that patch is analyzed, the straight line that detects due to Hough transformation not necessarily can be described route exactly, therefore need to jointly complete in conjunction with other algorithms the detection of road area.In general, the road surface color distribution is relatively fixing, the existing mode of asking for Color histogram distribution in processing according to image can identify out with road area, jointly limits road area in conjunction with the front by the detected Road information of Hough transformation method.
(1) patch analysis: be communicated with under environment four patch is analyzed.So-called four are communicated with, and refer to that four next-door neighbours' of direction up and down of a picture element point and this point is neighbouring relations, and upper left, lower-left, upper right, four of bottom rights point and this point do not belong to neighbouring relations.Patches information is abundanter, and vehicle detection and tracking are just more accurate.The essential information that the patch analysis is extracted mainly contains: the upper and lower, left and right coordinate of patch.Utilize this four coordinates, can calculate easily the center point coordinate of patch and the rectangular area of patch.
(2) center point coordinate of patch
Center point coordinate can reflect the position of vehicle roughly, is one of Main Basis of vehicle detection and tracking.
(3) plaque area
Plaque area comprises real area and rectangular area.Utilize area information can remove the impact of some noises.Further accurately obtain the position of patch.
(4) the RGB information of patch
Determine the position of patch in binary map after, find corresponding patch position in the RGB figure, then add up RGB information.RGB information is also one of Main Basis of vehicle detection and tracking.By statistics RGB information, can distinguish comparatively exactly target object and surrounding environment, become one of important step that detects stationary vehicle.
(5) patch number
In whole binary map, the number of all patches.Obtaining the patch number is convenient to the patch in whole image is unified to process.
By the patch analysis, utilize the RGB colouring information, can tentatively determine to realize the detection of stationary vehicle in the position of stationary vehicle.
Road traffic parameter is extracted
After road target is detected, can extract relevant parameter and come the visual evaluation Traffic Information.
1, the travel direction of vehicle
Determine the travel direction of vehicle according to the magnitude relationship of vehicle movement vector and background dot motion vector (being that ground stationary object is with respect to the motion vector of aircraft).The magnitude relationship that can obtain motion vector according to the observation to video is as follows:
Vector
The reverse driving vehicle＞vector
Background dot＞vector
Driving vehicle in the same way(1)
At first, in resulting unique point, background dot is maximum, so just can obtain the vector background dot, and then according to following formula with the motion vector of each car and vector background dot compare the travel direction of determining vehicle be with aircraft in the same way or reverse with aircraft.If the motion vector of vehicle is greater than the background dot motion vector, vehicle and aircraft reverse driving so; If the motion vector of vehicle is less than the background dot motion vector, vehicle and aircraft travel in the same way so.
2, the extraction of toy vehicle velocity value
Because the flying speed of aircraft is given, and the flying speed (given) of background characteristics motion of point vector ∝ aircraft, vehicle characteristics motion of point vector ∝ vehicle is with respect to the relative velocity of aircraft simultaneously, like this, just the relative velocity that can the proportion of utilization relation draws aircraft, the recycling following formula is extrapolated the actual travel speed of vehicle:
V
Car=V
Relatively-V
Aircraft(vehicle and aircraft reverse driving) (2)
V
Car=V
Relatively+ V
Aircraft(vehicle and aircraft travel in the same way) (3)
3, the extraction of vehicle
Length of wagon:
length＝(y
max-y
min)*cosθ*(V
plane*0.04/vector
background) (4)
Y wherein
max-y
minBe the displacement of vehicle on the aircraft flight direction, θ is the angle of aircraft flight direction and vehicle heading, V
PlaneThe flying speed of aircraft, vector
BackgroundBe the speed of related movement of background dot on image, 0.04 is interval time between every frame.
Following formula is to utilize known air speed equally, the actual range of aircraft flight and the pixel unit of image between adjacent two frames is mapped, thereby estimates the length of vehicle body.
Above process is all according to Aerial Images, utilizes computer image technology to process, and extracts by distinct methods for different parameters and realizes, finally obtains the required parameters of native system, comprises vehicle heading, speed and length of wagon.
Transport information based on Aerial Images is obtained system
This system processes by video that unmanned plane is taken photo by plane, and can obtain the traffic information on the highway section of taking photo by plane, and the information such as the world coordinates of each car, travel speed.On the basis that transport information is obtained, by analysis-by-synthesis and the processing to this information, as the statistics to number of vehicles, to measurement of the speed of a motor vehicle etc., draw the comprehensive traffic information on road surface, this area and detailed traffic parameter.
Owing to having adopted such scheme, the present invention has following characteristics: not only can be applied in the analysis of road conditions under the general magnitude of traffic flow, can carry out the traffic analysis for the sparse highway section in west area especially, obtain required traffic parameter and comprehensive road condition.And take photo by plane because the gained image derives from unmanned plane, the present invention more can adapt to the variation that gathers image.
Description of drawings
Fig. 1 is the schematic flow sheet of a kind of embodiment of the inventive method.
Fig. 2 is the moving target feature detection tracking process flow diagram of a kind of embodiment of the inventive method.
Fig. 3 is the static target detection method process flow diagram of a kind of embodiment of the inventive method.
Embodiment
The present invention is further illustrated below in conjunction with the accompanying drawing illustrated embodiment.
Native system is taken photo by plane on unmanned plane by camera and is obtained video, and video is correlated with obtains the input video that needs after pre-service.As shown in Figure 1, be the process that gathers video.Next, need to carry out a series of relevant treatment to obtain required content to video.
At first, video is cut into frame, each frame is equivalent to piece image, then processes for each width image.The below sets forth respectively the distinct methods for moving object detection and static target detection.
For moving target, the first step need to obtain its unique point.For each two field picture, as Fig. 2, utilize the KLT algorithm that image is scanned and obtain unique point, at first the KLT operator calculates eigenvalue λ 1, the λ 2 of the δ matrix of each unique point, if min is (λ
1, λ
2)＞H (threshold value) (in general threshold value is empirical value) this point is the validity feature point.The δ defined matrix is
In formula: I
xBe single order x directional derivative, I
yBe single order y directional derivative.
Along with the movement of camera, the intensity of image is occuring to change in the mode of complexity.If it is enough fast that video camera is caught the speed of image, so for consecutive frame, due to the various similaritys that affect the grey scale change factor, grey scale change in regional area is extremely similar, therefore can think, there is the displacement along X and Y-direction between the regional area of adjacent two frames, Here it is so-called two dimensional character translational motion model.This means that certain the unique point X=(x, y) on t time chart picture has moved to X '=(x-dx, y-dy) constantly at t+1, d=(dx wherein, dy) be the translation motion parameter vector of two dimensional character, the gray-scale value of this unique point is approximately equalised before and after motion, namely
J(X)＝I(X-d)+n(X) (6)
Wherein: J (X)=I (X, t+1) is the gray-scale value of t+1 unique point X constantly, and I (X-d)=I (X-d, t) is the t gray-scale value of this unique point constantly, and n (X) is corresponding noise.
Obviously need to select suitable kinematic parameter vector d to make the residual error that following double integral obtains in certain the characteristic window W around unique point X minimum
ε＝∫
W(I(X-d)-J(X))
2ωdX (7)
In formula, ω is the weighted equation to different picture elements in characteristic area.If in the smaller situation of the motion between adjacent two frames, I (X-d) can be carried out the single order Taylor expansion at the X point
I(X-d)＝J(X)-g·d (8)
G is gradient vector, so formula (7) can be write as following form again
ε＝∫
W(I(X-d)-J(X))
2ωdX＝∫
W(h-g·d)
2ωdX (9)
H=I (X)-J (X) wherein.Can find out, residual error is the quadratic equation of translation vector d, and this optimization problem can obtain the solution of closed form.In order to make residual error minimum, to formula (9) equal sign both sides, ask it to the first order derivative of d, obtain
∫
W(h-g·d)gωdX＝0 (10)
Due to (gd) g=(gg
T) d, and hypothesis d is constant in the characteristic window zone, therefore obtains
d×∫
W(gg
T)ωdX＝∫
WhgωdX (11)
Following formula is the basic calculating step during signature tracking calculates, can calculate it along the gradient of X and Y-direction for all pixels in characteristic window, therefore can obtain real symmetric intersection gradient matrix G, can calculate between two frames gray scale difference and obtain vectorial e for all pixels in characteristic window simultaneously.So just can calculate the value of kinematic parameter d.After obtaining d, moving characteristic window, then repeat above process, until d is less than certain threshold value, this shows the characteristic window of adjacent two frames, and the match is successful, with in the repetitive process of front, each is taken turns the d that obtains and all adds up and just obtain final translation motion parameter.
Second step for feature points clustering, accurately moving target is separated, needs to adopt effective clustering algorithm.What traditional clustering algorithm application was more is the k-means algorithm.
The course of work of k-means algorithm is described as follows: at first select arbitrarily k object as initial cluster center from n data object; And for other object of be left, according to the similarity (distance) of they and these cluster centres, respectively they are distributed to (cluster centre representative) cluster the most similar to it; And then calculate the cluster centre (average of all objects in this cluster) of each new cluster that obtains; Constantly repeat this process until the canonical measure function begins convergence.Generally all adopt mean square deviation to have following characteristics as .k cluster of canonical measure function: each cluster itself is compact as much as possible, and separates as much as possible between each cluster.
K-means has its shortcoming: the size of generation class differs can be very not large, very sensitive for dirty data.Based on this point, a kind of improved algorithm has been proposed: the k-medoids method.
Object of K-medoids algorithm picks is called the effect that medoid replaces top center, and such a medoid has just identified this class.Step:
(1) choose arbitrarily K object as medoids (O from unique point
1, O
2... O
i... O
k).
Below circulate:
(2) object of remainder is assigned in each class gone (according to the principle the most close with medoid);
(3) for each class (O
i) in, order is chosen an O
r, calculate and use O
rReplace O
iAfter consumption-E (O
r).Select that O of E minimum
rReplace O
iK medoids just changed like this, below just forward again (2) to.
(4) circulate until K medoids is fixed up like this.
This algorithm is insensitive for dirty data and abnormal data, but calculated amount is obviously large than K average, generally is only suitable for small data quantity.
Density-based k-medoids algorithm is in cluster process, does not need to input the number of cluster, but according to the distance between data, adjacent class is merged.
The 3rd step, the extraction of road traffic parameter.
1) extraction of direction of motion.The target motion vectors that is obtained by the front and background dot can draw direction of motion with respect to the motion vector of aircraft according to (1) formula.
2) extraction of toy vehicle velocity value.The relative velocity that is drawn by motion vector and the air speed of acquisition from take photo by plane can draw car speed according to formula (2) and formula (3).
3) vehicle is extracted.Can obtain Vehicle length according to coordinate and the motion vector of respective point by formula (4), and classify and extract vehicle.
For stationary vehicle, need to obtain the positional information of vehicle target.
As shown in Figure 3, at first carry out road area and select, in general, the road surface color distribution is relatively fixing, according to Color histogram distribution, road area can be identified out, jointly limits road area in conjunction with the detected Road information in front.
Secondly, adopt the method for patch analysis to determine static target position (coordinate).
For the patch analytical approach, need altogether twice sweep, scanning for the first time from the bottom up, scanning from left to right, scan pixel value and be 0 point, be left intact.Be 255 point when scanning first pixel value, it be labeled as " 1 ", and record.Then scanning four neighbours of this point, is 255 point if pixel value is arranged, and it is labeled as the mark " 1 " identical with this point, represents that they belong to same patch " 1 ".Record at last the coordinate figure of this point, and record " No. 1 patch has 1 pixel at present ".Then continue scanning, the step above repeating.
What scanning was mainly completed for the second time is the mark correction.
When merging pixel, the statistics of information has also been completed in current scanning, comprises the area of patch, the coordinate of upper and lower, left and right, center point coordinate, RGB information etc.
All extract required each traffic parameter, completion system function according to preceding method after detecting and finishing.
Native system can obtain traffic parameter and the traffic state informations such as the magnitude of traffic flow, the speed of a motor vehicle, vehicle on a certain highway section effectively, it is a kind of comprehensive system, comprise video input apparatus, video analysis instrument (software) and result display apparatus (computing machine), process that by systematic analysis result is shown, information needed is provided for related personnel or department.
The above-mentioned description to embodiment is can understand and apply the invention for ease of those skilled in the art.The person skilled in the art obviously can easily make various modifications to these embodiment, and in the General Principle of this explanation is applied to other embodiment and needn't pass through performing creative labour.Therefore, the invention is not restricted to the embodiment here, those skilled in the art should be within protection scope of the present invention for improvement and modification that the present invention makes according to announcement of the present invention.
Claims (4)
1. the method for acquiring traffic information based on Aerial Images, is characterized in that: analyze according to Aerial Images, detect dynamic object and static object, extract road traffic parameter; Described extraction road traffic parameter comprises extracts vehicle heading, car speed and length of wagon;
Wherein, the dynamic object detection method is for adopting the KLT algorithm to obtain the kinematic parameter of some unique points and dynamic object, determine the direction of motion vector by the road area direction projection, adopt the k-medoids algorithm with some feature points clusterings, thereby by unique point, dynamic object is separated;
Static object detection method is for adopting road area restriction and patch analytical approach to limit road area and static object;
Described KLT algorithm is under the similar condition of the characteristic area gray scale between the Aerial Images consecutive frame, choose a large amount of unique points in grayscale image sequence, obtain characteristic point position thereby carry out the two dimensional character tracking, solve accordingly the method for two dimensional character kinematic parameter; Simultaneously, the projecting direction according to motion vector extracts road direction, and utilizes Hough transformation to detect Road;
Described k-medoids algorithm comprises the following steps:
(1) choose arbitrarily K object as medoids(O from unique point
1, O
2... O
iO
k);
(2) object of remainder is assigned in each class according to the principle the most close with medoid gone;
(3) for each class (O
i) in, order is chosen an O
r, calculate and use O
rReplace O
iAfter consumption E(O
r), that O of selection E minimum
rReplace O
i
(4) return to step (2) cycle calculations, until K medoids is fixed up;
Wherein, described road area restriction is to adopt the mode of Color histogram distribution with the road area identification process, changes detected Road in conjunction with Hough and jointly limits road area;
Described patch analysis is RGB information and the patch information of number of the center point coordinate that specifically obtains patch, plaque area, patch, the position of the center point coordinate reflection static object of patch; Plaque area is used for removing noise; The RGB information of patch is after the image that obtains of taking photo by plane carries out binary conversion treatment, determines the position of patch in this image, then finds corresponding patch position in the former Aerial Images, and statistics RGB information is to distinguish target object and surrounding environment; The patch number is used for unified processing of patch to whole Aerial Images.
2. the method for acquiring traffic information based on Aerial Images as claimed in claim 1, it is characterized in that: described vehicle heading, to determine the travel direction of vehicle according to the magnitude relationship of vehicle movement vector and background dot motion vector, described vehicle movement vector refers to that dynamic object is with respect to the motion vector of aircraft, described background dot motion vector refers to static object with respect to the motion vector of aircraft, if the vehicle movement vector is greater than the background dot motion vector; Vehicle and aircraft reverse driving so; If the vehicle movement vector is less than the background dot motion vector, vehicle and aircraft travel in the same way so.
3. the method for acquiring traffic information based on Aerial Images as claimed in claim 1, it is characterized in that: described car speed calculates according to following formula,
When vehicle and aircraft reverse driving, V
Car=V
Relatively-V
Aircraft
When vehicle and aircraft travel in the same way, V
Car=V
Relatively+ V
Aircraft
Wherein, V
CarRefer to the speed of dynamic object, V
RelativelyRefer to the speed of static object, V
AircraftRefer to the speed of aircraft.
4. the method for acquiring traffic information based on Aerial Images as claimed in claim 1, it is characterized in that: described length of wagon is according to length=(y
max-y
min) * cos θ * (V
Plane* 0.04/vector
Background) calculating, wherein y
max-y
minBe the displacement of vehicle on the aircraft flight direction, θ is the angle of aircraft flight direction and vehicle heading, V
PlaneThe flying speed of aircraft, vector
BackgroundBe the speed of related movement of background dot on image, 0.04 is interval time between every frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105888800A CN102073846B (en) | 2010-12-15 | 2010-12-15 | Method for acquiring traffic information based on aerial images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105888800A CN102073846B (en) | 2010-12-15 | 2010-12-15 | Method for acquiring traffic information based on aerial images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102073846A CN102073846A (en) | 2011-05-25 |
| CN102073846B true CN102073846B (en) | 2013-06-05 |
Family
ID=44032380
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010105888800A Expired - Fee Related CN102073846B (en) | 2010-12-15 | 2010-12-15 | Method for acquiring traffic information based on aerial images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102073846B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015134453A1 (en) * | 2014-03-03 | 2015-09-11 | Inrix Inc. | Estimating traveler volume by evaluating aerial images |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103366555B (en) * | 2013-07-01 | 2015-06-17 | 中国人民解放军第三军医大学第三附属医院 | Aerial image-based traffic accident scene image rapid generation method and system |
| CN103413444B (en) * | 2013-08-26 | 2015-08-19 | 深圳市川大智胜科技发展有限公司 | A kind of traffic flow based on unmanned plane HD video is investigated method |
| US9175966B2 (en) * | 2013-10-15 | 2015-11-03 | Ford Global Technologies, Llc | Remote vehicle monitoring |
| BR102013033041B1 (en) * | 2013-12-20 | 2022-02-01 | Perkons S/A | System and method for monitoring and enforcing traffic and recording traffic violations and corresponding unmanned aerial vehicle |
| CN104035985B (en) * | 2014-05-30 | 2017-07-07 | 同济大学 | A kind of method for digging towards Fundamental Geographic Information System abnormal data |
| CN105046952B (en) * | 2015-06-17 | 2017-12-08 | 天津职业技术师范大学 | The sparse air-ground associated detecting method of road traffic accident |
| CN107871393B (en) * | 2016-09-27 | 2021-06-04 | 千寻位置网络有限公司 | Method and system for automatically judging whether vehicle occupies prohibited lane in case of violation of regulations |
| CN107038411A (en) * | 2017-02-26 | 2017-08-11 | 北京市交通运行监测调度中心 | A kind of Roadside Parking behavior precise recognition method based on vehicle movement track in video |
| CN107301369A (en) * | 2017-09-04 | 2017-10-27 | 南京航空航天大学 | Road traffic congestion analysis method based on Aerial Images |
| CN108009590A (en) * | 2017-12-13 | 2018-05-08 | 广东机场白云信息科技有限公司 | Flight information display system and its terminal equipment failure display methods |
| CN108564787A (en) * | 2018-05-31 | 2018-09-21 | 北京理工大学 | Traffic observation procedure, system and equipment based on Floating Car method |
| CN111382784B (en) * | 2020-03-04 | 2021-11-26 | 厦门星纵智能科技有限公司 | Moving target tracking method |
| CN114913469B (en) * | 2022-07-11 | 2022-11-22 | 浙江大华技术股份有限公司 | Method for establishing vehicle length estimation model, terminal equipment and storage medium |
| CN116612390B (en) * | 2023-07-21 | 2023-10-03 | 山东鑫邦建设集团有限公司 | Information management system for constructional engineering |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101368828A (en) * | 2008-10-15 | 2009-02-18 | 同济大学 | Blind man navigation method and system based on computer vision |
| CN101770581A (en) * | 2010-01-08 | 2010-07-07 | 西安电子科技大学 | Semi-automatic detecting method for road centerline in high-resolution city remote sensing image |
-
2010
- 2010-12-15 CN CN2010105888800A patent/CN102073846B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101368828A (en) * | 2008-10-15 | 2009-02-18 | 同济大学 | Blind man navigation method and system based on computer vision |
| CN101770581A (en) * | 2010-01-08 | 2010-07-07 | 西安电子科技大学 | Semi-automatic detecting method for road centerline in high-resolution city remote sensing image |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015134453A1 (en) * | 2014-03-03 | 2015-09-11 | Inrix Inc. | Estimating traveler volume by evaluating aerial images |
| EP3114663A4 (en) * | 2014-03-03 | 2018-03-14 | Inrix, Inc. | Estimating traveler volume by evaluating aerial images |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102073846A (en) | 2011-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102073846B (en) | Method for acquiring traffic information based on aerial images | |
| Spencer Jr et al. | Advances in computer vision-based civil infrastructure inspection and monitoring | |
| Zhu et al. | Pavement distress detection using convolutional neural networks with images captured via UAV | |
| CN110175576B (en) | Driving vehicle visual detection method combining laser point cloud data | |
| Varadharajan et al. | Vision for road inspection | |
| Zakeri et al. | Image based techniques for crack detection, classification and quantification in asphalt pavement: a review | |
| Radopoulou et al. | Patch detection for pavement assessment | |
| Zheng et al. | A novel vehicle detection method with high resolution highway aerial image | |
| Li et al. | Automatic crack recognition for concrete bridges using a fully convolutional neural network and naive Bayes data fusion based on a visual detection system | |
| CN110619279B (en) | Road traffic sign instance segmentation method based on tracking | |
| CN100544446C (en) | The real time movement detection method that is used for video monitoring | |
| Azevedo et al. | Automatic vehicle trajectory extraction by aerial remote sensing | |
| CN107463890B (en) | A kind of Foregut fermenters and tracking based on monocular forward sight camera | |
| CN103324936B (en) | A kind of vehicle lower boundary detection method based on Multi-sensor Fusion | |
| CN104463249B (en) | A kind of remote sensing images airfield detection method based on Weakly supervised learning framework | |
| CN105373135A (en) | Method and system for guiding airplane docking and identifying airplane type based on machine vision | |
| CN110910350B (en) | Nut loosening detection method for wind power tower cylinder | |
| CN108052904B (en) | Method and device for acquiring lane line | |
| CN103400151A (en) | Optical remote-sensing image, GIS automatic registration and water body extraction integrated method | |
| CN107767400A (en) | Remote sensing images sequence moving target detection method based on stratification significance analysis | |
| CN102663778B (en) | A kind of method for tracking target based on multi-view point video and system | |
| CN107808524A (en) | A kind of intersection vehicle checking method based on unmanned plane | |
| Wei et al. | Damage inspection for road markings based on images with hierarchical semantic segmentation strategy and dynamic homography estimation | |
| CN114973116A (en) | Method and system for detecting foreign matters embedded into airport runway at night by self-attention feature | |
| Hassan et al. | Detecting patches on road pavement images acquired with 3D laser sensors using object detection and deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130605 Termination date: 20201215 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |