CN103413444B - A kind of traffic flow based on unmanned plane HD video is investigated method - Google Patents
A kind of traffic flow based on unmanned plane HD video is investigated method Download PDFInfo
- Publication number
- CN103413444B CN103413444B CN201310375004.3A CN201310375004A CN103413444B CN 103413444 B CN103413444 B CN 103413444B CN 201310375004 A CN201310375004 A CN 201310375004A CN 103413444 B CN103413444 B CN 103413444B
- Authority
- CN
- China
- Prior art keywords
- traffic
- image
- video
- unmanned plane
- volume
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a kind of traffic flow based on unmanned plane HD video to investigate method, comprise the following steps: video acquisition step, unmanned plane is hovered over above selected urban road intersection and take HD video; Steady picture pre-treatment step, copies HD video and carries out the pre-service of steady picture, exporting sequence of pictures; Detect tracing step, the sequence of pictures after steady picture is carried out detection to moving object and tracking as derived data; Analytic statistics step, carries out wagon flow quantitative analysis and statistics to the Target id tracked and the current coordinates of motion; And, export step, the statistics of vehicle flowrate is transferred to Client GUI and shows, generate data and form.The present invention can obtain multiple high-precision traffic data, remarkable minimizing field investigation workload, do not affect traffic, there is the statistics of the whole magnitudes of traffic flow on direction, 12, crossing, for the improvement of blocking up such as intersection signal timing designing, traffic channelling improvement provide Data support.
Description
Technical field
The present invention relates to a kind of traffic flow to investigate method, particularly relate to a kind of traffic flow based on unmanned plane HD video and to investigate method.
Background technology
Vehicle identification and tracking are important contents in intelligent transport technology, have broad application prospects in future traffic information collection and traffic control and management; Existing vehicle identification and tracking technique are mainly studied based on the video of fixing camera, but fixing camera is limited by mobile difficulty, the shortcomings such as limited coverage area, greatly constrain its effect in urban transportation monitoring and play.And existing traditional traffic stream characteristics investigation mainly adopts the method investigated respectively the volume of traffic, the speed of a motor vehicle, density etc., not only on fact-finding process and data processing, need a large amount of man power and materials, but also there will be the problems such as interference traffic to a certain extent, investigation sample amount is little, fact-finding process can not reproduce.
Summary of the invention
Technical matters to be solved by this invention needs to provide one not only can significantly reduce field investigation workload, but also have and do not affect traffic, Data acquisition and issuance processing procedure can repeatedly be carried out, and the traffic flow with long preservation period of traffic live material is investigated method.
To this, the invention provides a kind of traffic flow based on unmanned plane HD video and to investigate method, comprise the following steps:
Video acquisition step, hovers over unmanned plane above selected urban road intersection, and takes HD video;
Steady picture pre-treatment step, copies HD video and carries out the pre-service of steady picture, exporting sequence of pictures;
Detect tracing step, the sequence of pictures after steady picture is carried out detection to moving object and tracking as derived data;
Analytic statistics step, carries out wagon flow quantitative analysis and statistics to the Target id tracked and the current coordinates of motion; And,
Export step, the statistics of vehicle flowrate is transferred to Client GUI and shows, and generate data and form.
The invention belongs to digital image processing techniques field, be specifically related to a kind of traffic flow based on unmanned plane HD video by electronic steady image and visual analysis to investigate method, this traffic flow investigates method first by surely carrying out pre-service as device to the HD video that unmanned plane is taken, then output to target acquistion software and carry out moving object segmentation and tracking, again result is transferred to trajectory analysis software and carries out vehicle flowrate, finally output to graphical interfaces to show, data and form are passed judgment in the quantification generating traffic flow.
Unmanned plane (UAV) progresses into traffic engineering field as a kind of emerging information acquisition means, unmanned plane can carry out motor-driven, flexible and large-scale traffic information collection by carrying high-resolution picture pick-up device, information acquisition in can advancing according to certain path like this, can hover again obtains traffic video above observation area.But unmanned plane (UAV) gathers transport information also its intrinsic difficulty, as: video shake, due to video sensor move along with the high-speed motion of unmanned plane, the feature such as the complicacy of background and the diversity of moving target information in video sequence image, these features make the problem of processing target detecting and tracking become more difficult.
The present invention with the traffic flow of crossing for research object, the traffic flow video of crossing is gathered by the mode of hovering directly over unmanned plane (UAV), and the video that unmanned plane (UAV) gathers is analyzed, identify the vehicle individual in crossing, the speed of a motor vehicle and follow the tracks of the motion process of vehicle; By the feature of research crossing hovering unmanned plane (UAV) video, design and can overcome video jitter, translation, rotate and the vehicle identification of the adverse effect such as to fluctuate and tracking processing method, and then achieve effective extraction of transport information in crossing.
Compared with prior art, the means of the present invention's application video encoder server carry out investigation to traffic stream characteristics and process, Data acquisition and issuance process is carried out by the traffic fact of recording video camera, the traffic data that multiple precision is very high can be obtained, not only can significantly reduce field investigation workload, but also have and do not affect traffic, Data acquisition and issuance processing procedure can repeatedly be carried out, and traffic live material can be preserved for a long time; On this basis, the present invention compares other similar section formula video analysis traffic flow investigating systems and compares, overcome unmanned plane (UAV) and gather the intrinsic difficulty in transport information, and there is the statistical function of the whole magnitudes of traffic flow on direction, 12, crossing, can the rule of blocking up and being formed and dissipating be summed up, for the control measures that block up such as intersection signal timing designing, traffic channelling improvement provide Data support.
Further improvement of the present invention is, describedly surely comprise following sub-step as pre-treatment step: step one, translation, Rotation and Zoom are carried out to current frame image, makes itself and previous frame image there is no difference, and the motion-vector of current frame image and previous frame image is write journal file; Step 2, carries out translation, Rotation and Zoom according to journal file to current frame image.Preferably, the described steady step one as pre-treatment step is: first, searches the violent shake of current frame image and previous frame image, downscaled images displacement calculating amount; Then, the image reduced is carried out respectively to the amplification of twice and four times, the displacement size of meticulous these image blocks of adjustment is until find the motion-vector of original image.Search motion-vector by reducing two field picture supporting and amplify, the shake of video can be eliminated quickly and accurately.
Further improvement of the present invention is, described detection tracing step comprises following sub-step:
Noise reduction process sub-step, adopts median filtering method to carry out noise reduction process to gray level image;
Binary conversion treatment sub-step, adopts threshold segmentation method to carry out binary conversion treatment to image;
Vehicle detection sub-step, adopts background subtraction to carry out vehicle detection; And,
Vehicle tracking sub-step, by comprising vehicle identifiers, the agglomerate of position and size follows the tracks of vehicle.
Further improvement of the present invention is, described noise reduction process sub-step first carries out the calculating of maximal value, intermediate value and minimum value to each column data in sequence of pictures, then get the output pixel value of mean as filter result to maximal value, intermediate value and minimum value.
For ease of illustrating, for the video in window of 3 x 3, be P0, P1 by each pixel definition in 3 x 3 windows ... P8; To each column count maximal value, intermediate value and minimum value in window, and the output pixel value of mean as filter result is got to maximal value, intermediate value and minimum value, this median calculation only needs 17 times to compare, about 2 times are decreased compared with traditional algorithm, can obtain a result rapidly, be applicable to real-time process.
Further improvement of the present invention is, described binary conversion treatment sub-step presets threshold value T, according to
image is divided into two pixel groups being greater than T He being less than T.The present invention sets a certain threshold value T(OSTU maximum variance between clusters), this threshold value T can carry out setting and changing as the case may be, the data of image can be divided into two parts by threshold value T: be greater than the pixel group of T and be less than the pixel group of T, be convenient to follow-uply carry out visual analysis to image.
Further improvement of the present invention is, described vehicle detection sub-step first chooses at least piece image the average image as a setting in background, then the present frame of successive image sequence and background image subtraction, carries out background cancellation; When obtained pixel count be greater than preset threshold value T time, then judge there is moving target in observed scene.
First choose the width in background or a few width images the average image as a setting, then the present frame of successive image sequence and background image subtraction, carry out background cancellation; If the pixel count obtained is greater than a certain threshold value, then can judges there is moving target in observed scene, be formulated as follows:
,
; Wherein,
for present frame,
for background image.
Background extracting and renewal adopt Gaussian mixtures background model, and update rule is as follows:
.Wherein,
represent the threshold value of noise; X represents current frame pixel value; α represents the learning rate of model, i.e. renewal rate; Because the renewal of background model also not exclusively depends on current pixel value, and have correlativity with front frame, after introducing α, background model can keep relative stability in long-time.
Further improvement of the present invention is, described vehicle tracking sub-step is: carry out Iamge Segmentation to candidate pixel, judges that the pixel of this segmentation image belongs to target, background or other regions, the pixel with same characteristic features is assembled a region; Use contour area to identify vehicle as characteristic quantity, the same moving target of images different in video flowing sequence is connected, indicates by oval frame; Carry out Kalman's estimation, in order to provide speed and the accuracy of new Iamge Segmentation region detection, circulate according to this.
Further improvement of the present invention is, described analytic statistics step is Track Pick-up and track aftertreatment, comprise the analysis to target data number, event data number and timestamp, described target data number comprises the ID of target, barycenter image coordinate, image coordinate width, image coordinate height, speed, type, track, place and region; Described event data number comprises mark, Target id number, event, track, lights state, speed and event additional data.
Further improvement of the present invention is, the form that described output step exports comprises short-term report and long-term observation report, the data of described short-term report comprise the 5min volume of traffic, daytime the 12h volume of traffic, peak hour traffic, peak hour flow ratio, peak hour factor, road direction distribution coefficient, the busy direction volume of traffic account for ratio and right-hand rotation, craspedodrome and the left-hand bend wagon flow ratio of the total volume of traffic; The data of described long-term observation report comprise annual average daily traffic, all average daily traffic volumes, monthly average daily traffic, the annual moon volume of traffic, year the highest hourly traffic volume, the 30th maximum annual hourly volume, the 30th volume of traffic coefficient, the moon volume of traffic variation factor and all daily traffic volume variation factors.
Compared with prior art, beneficial effect of the present invention is, by HD video to vehicle and detect, and then the survey and analysis process realized traffic flow, the traffic data that multiple precision is very high can be obtained, not only can significantly reduce field investigation workload, but also have and do not affect traffic, Data acquisition and issuance processing procedure can repeatedly be carried out, and traffic live material can be preserved for a long time; On this basis, present invention overcomes unmanned plane (UAV) gathering the video jitter in transport information, technical matters that video sensor moves along with the high-speed motion of unmanned plane, background is complicated in video sequence image, moving target information is various etc. brings, there is the statistical function of the whole magnitudes of traffic flow on direction, 12, crossing, can the rule of blocking up and being formed and dissipating be summed up, for the control measures that block up such as intersection signal timing designing, traffic channelling improvement provide Data support.
Accompanying drawing explanation
Fig. 1 is the workflow schematic diagram of an embodiment of the present invention;
Fig. 2 is the workflow schematic diagram of the background subtraction of the vehicle detection sub-step of the another kind of embodiment of the present invention;
Fig. 3 is the renewal schematic flow sheet of the background model of the vehicle detection sub-step of the another kind of embodiment of the present invention;
Fig. 4 is the previous moment of the another kind of embodiment of the present invention and the variance prediction of current time and revises schematic diagram;
Fig. 5 is the workflow schematic diagram of the vehicle tracking sub-step of the another kind of embodiment of the present invention.
Embodiment
Below in conjunction with accompanying drawing, preferably embodiment of the present invention is described in further detail.
Embodiment 1:
This example provides a kind of traffic flow based on unmanned plane HD video to investigate method, comprises the following steps:
Video acquisition step, hovers over unmanned plane above selected urban road intersection, and takes HD video;
Steady picture pre-treatment step, is copied to HD video on PC and carries out the pre-service of steady picture, export sequence of pictures;
Detect tracing step, the sequence of pictures after steady picture is carried out detection to moving object and tracking as derived data;
Analytic statistics step, carries out wagon flow quantitative analysis and statistics to the Target id tracked and the current coordinates of motion; And,
Export step, the statistics of vehicle flowrate is transferred to Client GUI and shows, and generate data and form.
The workflow schematic diagram of this example as shown in Figure 1, belong to digital image processing techniques field, be specifically related to a kind of traffic flow based on unmanned plane HD video by electronic steady image and visual analysis to investigate method, this traffic flow investigates method first by surely carrying out pre-service as device to the HD video that unmanned plane is taken, then output to target acquistion software and carry out moving object segmentation and tracking, again result is transferred to trajectory analysis software and carries out vehicle flowrate, finally output to graphical interfaces to show, data and form are passed judgment in the quantification generating traffic flow.
Unmanned plane (UAV) progresses into traffic engineering field as a kind of emerging information acquisition means, unmanned plane can carry out motor-driven, flexible and large-scale traffic information collection by carrying high-resolution picture pick-up device, information acquisition in can advancing according to certain path like this, can hover again obtains traffic video above observation area.But unmanned plane (UAV) gathers transport information also its intrinsic difficulty, as: video shake, due to video sensor move along with the high-speed motion of unmanned plane, the feature such as the complicacy of background and the diversity of moving target information in video sequence image, these features make the problem of processing target detecting and tracking become more difficult.
This example with the traffic flow of crossing for research object, the traffic flow video of crossing is gathered by the mode of hovering directly over unmanned plane (UAV), and the video that unmanned plane (UAV) gathers is analyzed, identify the vehicle individual in crossing, the speed of a motor vehicle and follow the tracks of the motion process of vehicle; By the feature of research crossing hovering unmanned plane (UAV) video, design and can overcome video jitter, translation, rotate and the vehicle identification of the adverse effect such as to fluctuate and tracking processing method, and then achieve effective extraction of transport information in crossing.
Compared with prior art, the means of this example application video encoder server carry out investigation to traffic stream characteristics and process, Data acquisition and issuance process is carried out by the traffic fact of recording video camera, the traffic data that multiple precision is very high can be obtained, not only can significantly reduce field investigation workload, but also have and do not affect traffic, Data acquisition and issuance processing procedure can repeatedly be carried out, and traffic live material can be preserved for a long time; On this basis, this example is compared other similar section formula video analysis traffic flow investigating systems and is compared, overcome unmanned plane (UAV) and gather the intrinsic difficulty in transport information, and there is the statistical function of the whole magnitudes of traffic flow on direction, 12, crossing, can the rule of blocking up and being formed and dissipating be summed up, for the control measures that block up such as intersection signal timing designing, traffic channelling improvement provide Data support.
Embodiment 2:
On the basis of embodiment 1, this example is described surely comprises step quickly as pre-treatment step: step one, carries out translation, Rotation and Zoom to current frame image, makes itself and previous frame image not have difference.First, search violent shake, vast scale downscaled images displacement calculating amount; Then, the image reduced is carried out to the amplification of twice and four times, motion-vector, until find the motion-vector (motion vectors) of original image, is finally write journal file by the displacement size of meticulous these image blocks of adjustment.Step 2, carries out translation, Rotation and Zoom according to journal file to current frame image.
This routine described detection tracing step and analytic statistics step all belong to visual analysis, specifically comprise following double teacher.
Stage 1: adopting Fast median filtering algorithm to carry out noise reduction process to gray level image, for ease of illustrating, is P0, P1 by each pixel definition in 3 x 3 windows ... P8; To each column count maximal value, intermediate value and minimum value in window, processing procedure comprises: Max0=max [P0, P3, P6], Max1=max [P1, P4, P7], Max2=max [P2, P5, P8], Med0=med [P0, P3, P6], Med1=med [P1, P4, P7], Med2=med [P2, P5, P8], Min0=min [P0, P3, P6], Min1=min [P1, P4, P7], Min2=min [P2, P5, P8].
The processing procedure of the output pixel value Winmed of filter result is as follows: Maxmin=min [Max0, Max1, Max2], Medmed=med [Med0, Med1, Med2], Minmax=max [Min0, Min1, Min2], Winmed=med [Maxmin, Medmed, Minmax].Only need 17 times to compare by above-mentioned median calculation, decrease about 2 times compared with traditional algorithm, be applicable to real-time process.
Stage 2: adopt threshold segmentation method to carry out binary conversion treatment to image.Set a certain threshold value T(OSTU maximum variance between clusters), with this threshold value T, the data of image can be divided into two parts: be greater than the pixel group of T and be less than the pixel group of T, the formula that binary conversion treatment adopts is:
.
Stage 3: adopt background subtraction to carry out vehicle detection.First choose the width in background or a few width images the average image as a setting, then the present frame of successive image sequence and background image subtraction, carry out background cancellation; If the pixel count obtained is greater than a certain threshold value, then can judges there is moving target in observed scene, be formulated as follows:
,
; Wherein,
for present frame,
for background image.This stage 2 and stage 3 belong to detection tracing step, and Fig. 2 is the workflow schematic diagram of the background subtraction of vehicle detection sub-step.
Background extracting and renewal adopt Gaussian mixtures background model, and update rule is as follows:
.
wherein,
represent the threshold value of noise, x represents current frame pixel value, and α represents the learning rate of model, i.e. renewal rate.Because the renewal of background model also not exclusively depends on current pixel value, and have correlativity with front frame, after introducing α, background model can keep relative stability in long-time, and Fig. 3 is the renewal schematic flow sheet of the background model of the vehicle detection sub-step of this example.
Stage 4: based on the vehicle tracking of agglomerate (blob).Iamge Segmentation is carried out to candidate pixel, judges that this pixel belongs to target or belongs to background or belong to other regions, the pixel with same characteristic features is assembled a region; Use contour area to identify vehicle as characteristic quantity, the same moving target of images different in video flowing sequence is connected, indicates by oval frame; Carry out Kalman's estimation again, in order to provide speed and the accuracy of new Iamge Segmentation region detection, circulate according to this.Wherein, Fig. 4 be this example be applied to traffic flow investigate previous moment in system and current time variance prediction and revise schematic diagram; Fig. 5 is the workflow schematic diagram of the vehicle tracking sub-step of this example.
Stage 5: Track Pick-up and track aftertreatment.The disposal route of real time data and picture bag can realize by program below.
typedef struct DetailDataHeader
{
Int nJPGLen; //jpeg figure leaf length
Int nWidth, nHeight; // picture width, picture height
Int nObjCnt; // target data number
Int nEvtCnt; // event data number
Int nOrient; //Jpeg image direction
ULONG ulFrames; // visual frame number
_ TIME_STAMP stamp; // timestamp
}DetailDataHeader;
Can realize by program below the disposal route of target data.
typedef struct ObjData
{
DWORD dwFlag; // mark
Int nID; // Target id;
Int x, y; // target centroid image coordinate
Int width, height; The image coordinate width of // target and height
Float fWidth; // target developed width, world coordinates width after namely demarcating, unit rice
Float fSpeed; // target velocity m/s
Int type; // target type, 0-vehicle, 1 pedestrian, 2 abandons, 3 bicycles
Int nPlateType; // car plate type
Int nBelieve; // recognition confidence
Char szPlate [16]; // the number-plate number
Int nRoad; Track, // target place
DWORD EvtMask; The event mask that // target once occurred
Int nArea; // target region
}ObjData;
Can realize by program below the disposal route of event data.
typedef struct EvtData
{
DWORD dwFlag; // mark
Int nID; // Target id number
DWORD dwEvent; // event
Int nRoad; // track
DWORD dwTraficFlag; // lights state
Float dbVelocity; // speed m/s
LONGLONG llPicID; // associate 832 normal pictures
Char AppendData [16]; // event additional data
} EvtData;
As can be seen from above program, this routine described analytic statistics step is Track Pick-up and track aftertreatment, comprise the analysis to target data number, event data number and timestamp, described target data number comprises the ID of target, barycenter image coordinate, image coordinate width, image coordinate height, speed, type, track, place and region; Described event data number comprises mark, Target id number, event, track, lights state, speed and event additional data.
The form that this routine described output step exports comprises short-term report and long-term observation report, the data of described short-term report comprise the 5min volume of traffic, daytime the 12h volume of traffic, peak hour traffic, peak hour flow ratio, peak hour factor, road direction distribution coefficient, the busy direction volume of traffic account for ratio and right-hand rotation, craspedodrome and the left-hand bend wagon flow ratio of the total volume of traffic; The data of described long-term observation report comprise annual average daily traffic, all average daily traffic volumes, monthly average daily traffic, the annual moon volume of traffic, year the highest hourly traffic volume, the 30th maximum annual hourly volume, the 30th volume of traffic coefficient, the moon volume of traffic variation factor and all daily traffic volume variation factors.
Compared with prior art, the beneficial effect of this example is, by HD video to vehicle and detect, and then the survey and analysis process realized traffic flow, the traffic data that multiple precision is very high can be obtained, not only can significantly reduce field investigation workload, but also have and do not affect traffic, Data acquisition and issuance processing procedure can repeatedly be carried out, and traffic live material can be preserved for a long time; On this basis, present invention overcomes unmanned plane (UAV) gathering the video jitter in transport information, technical matters that video sensor moves along with the high-speed motion of unmanned plane, background is complicated in video sequence image, moving target information is various etc. brings, there is the statistical function of the whole magnitudes of traffic flow on direction, 12, crossing, can the rule of blocking up and being formed and dissipating be summed up, for the control measures that block up such as intersection signal timing designing, traffic channelling improvement provide Data support.
Above content is in conjunction with concrete preferred implementation further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, some simple deduction or replace can also be made, all should be considered as belonging to protection scope of the present invention.
Claims (7)
1. to investigate a method based on the traffic flow of unmanned plane HD video, it is characterized in that, comprise the following steps:
Video acquisition step, hovers over unmanned plane above selected urban road intersection, and takes HD video;
Steady picture pre-treatment step, copies HD video and carries out the pre-service of steady picture, exporting sequence of pictures;
Detect tracing step, the sequence of pictures after steady picture is carried out detection to moving object and tracking as derived data;
Analytic statistics step, carries out wagon flow quantitative analysis and statistics to the Target id tracked and the current coordinates of motion, obtains the whole magnitudes of traffic flow on direction, 12, crossing, so statistics and sum up crossing block up formed and dissipate rule;
And, export step, the statistics of vehicle flowrate is transferred to Client GUI and shows, and generate data and form;
Describedly surely comprise following sub-step as pre-treatment step: step one, translation, Rotation and Zoom are carried out to current frame image, make itself and previous frame image there is no difference, and by the motion-vector write journal file of current frame image and previous frame image; Step 2, carries out translation, Rotation and Zoom according to journal file to current frame image;
Described detection tracing step comprises following sub-step: noise reduction process sub-step, adopts median filtering method to carry out noise reduction process to gray level image; Binary conversion treatment sub-step, adopts threshold segmentation method to carry out binary conversion treatment to image; Vehicle detection sub-step, adopts background subtraction to carry out vehicle detection; And, vehicle tracking sub-step, by comprising vehicle identifiers, the agglomerate of position and size follows the tracks of vehicle;
In described vehicle detection sub-step, first choose at least piece image the average image as a setting in background, then the present frame of successive image sequence and background image subtraction, carry out background cancellation; When obtained pixel count be greater than preset threshold value T time, then judge there is moving target in observed scene, be formulated as follows: d=|F
k(x, y)-B
k(x, y) |,
Wherein, F
k(x, y) is present frame, B
k(x, y) is background image; Background extracting and renewal adopt Gaussian mixtures background model, and update rule is as follows:
Wherein,
represent the threshold value of noise; X represents current frame pixel value; α represents the learning rate of model.
2. the traffic flow based on unmanned plane HD video according to claim 1 is investigated method, it is characterized in that, the described steady step one as pre-treatment step is: first, searches the violent shake of current frame image and previous frame image, downscaled images displacement calculating amount; Then, the image reduced is carried out respectively to the amplification of twice and four times, the displacement size of meticulous these image blocks of adjustment is until find the motion-vector of original image.
3. the traffic flow based on unmanned plane HD video according to claim 1 and 2 is investigated method, it is characterized in that, described noise reduction process sub-step first carries out the calculating of maximal value, intermediate value and minimum value to each column data in sequence of pictures, then get the output pixel value of intermediate value as filter result to maximal value, intermediate value and minimum value.
4. the traffic flow based on unmanned plane HD video according to claim 1 and 2 is investigated method, and it is characterized in that, described binary conversion treatment sub-step presets threshold value T, according to
Image is divided into two pixel groups being greater than T He being less than T.
5. the traffic flow based on unmanned plane HD video according to claim 1 and 2 is investigated method, it is characterized in that, described vehicle tracking sub-step is: carry out Iamge Segmentation to candidate pixel, judge that the pixel of this segmentation image belongs to target, background or other regions, the pixel with same characteristic features is assembled a region; Use contour area to identify vehicle as characteristic quantity, the same moving target of images different in video flowing sequence is connected, indicates by oval frame; Carry out Kalman's estimation, in order to provide speed and the accuracy of new Iamge Segmentation region detection, circulate according to this.
6. the traffic flow based on unmanned plane HD video according to claim 1 and 2 is investigated method, it is characterized in that, described analytic statistics step is Track Pick-up and track aftertreatment, comprise the analysis to target data number, event data number and timestamp, described target data number comprises the ID of target, barycenter image coordinate, image coordinate width, image coordinate height, speed, type, track, place and region; Described event data number comprises mark, Target id number, event, track, lights state, speed and event additional data.
7. the traffic flow based on unmanned plane HD video according to claim 6 is investigated method, it is characterized in that, the form that described output step exports comprises short-term report and long-term observation report, the data of described short-term report comprise the 5min volume of traffic, daytime the 12h volume of traffic, peak hour traffic, peak hour flow ratio, peak hour factor, road direction distribution coefficient, the busy direction volume of traffic account for ratio and right-hand rotation, craspedodrome and the left-hand bend wagon flow ratio of the total volume of traffic; The data of described long-term observation report comprise annual average daily traffic, all average daily traffic volumes, monthly average daily traffic, the annual moon volume of traffic, year the highest hourly traffic volume, the 30th maximum annual hourly volume, the 30th volume of traffic coefficient, the moon volume of traffic variation factor and all daily traffic volume variation factors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310375004.3A CN103413444B (en) | 2013-08-26 | 2013-08-26 | A kind of traffic flow based on unmanned plane HD video is investigated method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310375004.3A CN103413444B (en) | 2013-08-26 | 2013-08-26 | A kind of traffic flow based on unmanned plane HD video is investigated method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103413444A CN103413444A (en) | 2013-11-27 |
| CN103413444B true CN103413444B (en) | 2015-08-19 |
Family
ID=49606449
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310375004.3A Active CN103413444B (en) | 2013-08-26 | 2013-08-26 | A kind of traffic flow based on unmanned plane HD video is investigated method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103413444B (en) |
Families Citing this family (38)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103632535B (en) * | 2013-11-19 | 2015-06-10 | 河海大学 | Judgment method for section pedestrian crossing signal lamp arrangement |
| 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 |
| CN103730015B (en) * | 2013-12-27 | 2016-01-20 | 株洲南车时代电气股份有限公司 | Intersection traffic flow detecting method and device |
| CN103778790B (en) * | 2014-01-13 | 2015-08-12 | 杭州电子科技大学 | A kind of vehicle flowrate square wave statistic law based on video sequence |
| CN104134222B (en) * | 2014-07-09 | 2017-02-15 | 郑州大学 | Traffic flow monitoring image detecting and tracking system and method based on multi-feature fusion |
| CN104157165B (en) * | 2014-08-11 | 2016-07-06 | 江苏恒创软件有限公司 | A kind of urban area motor vehicle parking planing method based on unmanned plane |
| CN104269054B (en) * | 2014-09-20 | 2015-10-28 | 东北石油大学 | Based on the road section traffic volume index estimating system that unmanned plane is measured |
| CN104240508B (en) * | 2014-09-20 | 2016-05-04 | 青岛橡胶谷知识产权有限公司 | For the unmanned plane detection system of section jam alarming |
| US10102586B1 (en) | 2015-04-30 | 2018-10-16 | Allstate Insurance Company | Enhanced unmanned aerial vehicles for damage inspection |
| US9505494B1 (en) | 2015-04-30 | 2016-11-29 | Allstate Insurance Company | Enhanced unmanned aerial vehicles for damage inspection |
| US11140326B2 (en) | 2015-05-22 | 2021-10-05 | The United States Of America, As Represented By The Secretary Of The Navy | Aerial video based point, distance, and velocity real-time measurement system |
| EP3347789B1 (en) * | 2015-09-11 | 2021-08-04 | SZ DJI Technology Co., Ltd. | Systems and methods for detecting and tracking movable objects |
| WO2017088096A1 (en) | 2015-11-23 | 2017-06-01 | 深圳市大疆创新科技有限公司 | Method and device for synchronizing and collecting data of flying apparatuses, and flying apparatus |
| CN105321342B (en) * | 2015-12-07 | 2017-11-21 | 北京航空航天大学 | A kind of intersection vehicles queue length detection method based on video of taking photo by plane |
| WO2017114496A1 (en) * | 2015-12-31 | 2017-07-06 | Wellen Sham | Facilitating location positioning service through a uav network |
| CN105528891A (en) * | 2016-01-13 | 2016-04-27 | 深圳市中盟科技有限公司 | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring |
| CN106056624A (en) * | 2016-05-24 | 2016-10-26 | 西北工业大学 | Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof |
| CN105957341A (en) * | 2016-05-30 | 2016-09-21 | 重庆大学 | Wide area traffic jam detection method based on unmanned plane airborne platform |
| CN106791439B (en) * | 2017-01-22 | 2019-09-10 | 陕西外号信息技术有限公司 | A kind of stabilization compensation method of optical label |
| CN107248296B (en) * | 2017-07-13 | 2020-04-24 | 南京航空航天大学 | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics |
| US10410055B2 (en) * | 2017-10-05 | 2019-09-10 | TuSimple | System and method for aerial video traffic analysis |
| CN109658688A (en) * | 2017-10-11 | 2019-04-19 | 深圳市哈工大交通电子技术有限公司 | The detection method and device of access connection traffic flow based on deep learning |
| CN107808524B (en) * | 2017-11-30 | 2021-09-24 | 中山大学 | Road intersection vehicle detection method based on unmanned aerial vehicle |
| CN108009494A (en) * | 2017-11-30 | 2018-05-08 | 中山大学 | A kind of intersection wireless vehicle tracking based on unmanned plane |
| CN110278717B (en) * | 2018-01-22 | 2022-08-02 | 深圳市大疆创新科技有限公司 | Method and device for controlling the flight of an aircraft |
| CN108364466B (en) * | 2018-02-11 | 2021-01-26 | 金陵科技学院 | Traffic flow statistical method based on unmanned aerial vehicle traffic video |
| CN108399745B (en) * | 2018-03-01 | 2020-10-16 | 北京航空航天大学合肥创新研究院 | Unmanned aerial vehicle-based time-interval urban road network state prediction method |
| CN108694829B (en) * | 2018-03-27 | 2021-08-13 | 西安科技大学 | Traffic flow identification monitoring network and method based on unmanned aerial vehicle group mobile platform |
| CN108320510B (en) * | 2018-04-03 | 2020-12-04 | 深圳市智绘科技有限公司 | Traffic information statistical method and system based on aerial video shot by unmanned aerial vehicle |
| CN108648181A (en) * | 2018-04-25 | 2018-10-12 | 佛山科学技术学院 | A kind of automobile quantity statistics method and its system based on particle filter algorithm |
| 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 |
| CN109003451B (en) * | 2018-07-04 | 2020-11-06 | 北京航空航天大学 | Intersection OD matrix estimation method based on low-altitude unmanned aerial vehicle |
| CN109118767A (en) * | 2018-09-05 | 2019-01-01 | 盐城骏拔汽车零部件有限公司 | Intelligent transportation system based on unmanned plane |
| EP3748602A1 (en) * | 2019-06-03 | 2020-12-09 | Sony Corporation | Monitoring vehicle movement for traffic risk mitigation |
| CN110889971B (en) * | 2019-11-08 | 2021-11-30 | 深圳市国佳智能软件有限公司 | Intelligent vehicle finding method and device for parking lot |
| CN112735164B (en) * | 2020-12-25 | 2022-08-05 | 北京智能车联产业创新中心有限公司 | Test data construction method and test method |
| CN113706932B (en) * | 2021-10-29 | 2022-02-25 | 天津云圣智能科技有限责任公司 | Road section safety warning method and device and unmanned aerial vehicle |
| CN117315934A (en) * | 2023-09-25 | 2023-12-29 | 阜阳交通能源投资有限公司 | Expressway traffic flow real-time monitoring and congestion prediction system based on unmanned aerial vehicle |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11120481A (en) * | 1997-10-17 | 1999-04-30 | Mitsubishi Heavy Ind Ltd | Traffic flow measuring device |
| CN101009021A (en) * | 2007-01-25 | 2007-08-01 | 复旦大学 | Video stabilizing method based on matching and tracking of characteristic |
| JP2009087316A (en) * | 2007-09-11 | 2009-04-23 | Hitachi Ltd | Traffic flow measurement system |
| CN101968924A (en) * | 2010-10-19 | 2011-02-09 | 北方工业大学 | High-resolution image-based real-time traffic flow detection method |
| CN102073846A (en) * | 2010-12-15 | 2011-05-25 | 同济大学 | Method for acquiring traffic information based on aerial images |
| CN102289940A (en) * | 2011-07-26 | 2011-12-21 | 西南交通大学 | Hybrid differential-based traffic flow detection method |
| CN102436738A (en) * | 2011-09-26 | 2012-05-02 | 同济大学 | Traffic monitoring device based on unmanned aerial vehicle (UAV) |
| CN102496276A (en) * | 2011-12-01 | 2012-06-13 | 青岛海信网络科技股份有限公司 | High efficiency vehicle detection method |
| CN102509287A (en) * | 2011-10-12 | 2012-06-20 | 西安理工大学 | Finding method for static target based on latitude and longitude positioning and image registration |
-
2013
- 2013-08-26 CN CN201310375004.3A patent/CN103413444B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11120481A (en) * | 1997-10-17 | 1999-04-30 | Mitsubishi Heavy Ind Ltd | Traffic flow measuring device |
| CN101009021A (en) * | 2007-01-25 | 2007-08-01 | 复旦大学 | Video stabilizing method based on matching and tracking of characteristic |
| JP2009087316A (en) * | 2007-09-11 | 2009-04-23 | Hitachi Ltd | Traffic flow measurement system |
| CN101968924A (en) * | 2010-10-19 | 2011-02-09 | 北方工业大学 | High-resolution image-based real-time traffic flow detection method |
| CN102073846A (en) * | 2010-12-15 | 2011-05-25 | 同济大学 | Method for acquiring traffic information based on aerial images |
| CN102289940A (en) * | 2011-07-26 | 2011-12-21 | 西南交通大学 | Hybrid differential-based traffic flow detection method |
| CN102436738A (en) * | 2011-09-26 | 2012-05-02 | 同济大学 | Traffic monitoring device based on unmanned aerial vehicle (UAV) |
| CN102509287A (en) * | 2011-10-12 | 2012-06-20 | 西安理工大学 | Finding method for static target based on latitude and longitude positioning and image registration |
| CN102496276A (en) * | 2011-12-01 | 2012-06-13 | 青岛海信网络科技股份有限公司 | High efficiency vehicle detection method |
Non-Patent Citations (1)
| Title |
|---|
| 基于OpenCV的视频道路车辆检测与跟踪;詹峰群;《中国优秀硕士学位论文全文数据库信息科技辑》;20100115(第01期);第14-16、31、41-48、63页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103413444A (en) | 2013-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103413444B (en) | A kind of traffic flow based on unmanned plane HD video is investigated method | |
| Uke et al. | Moving vehicle detection for measuring traffic count using opencv | |
| CN100573618C (en) | A kind of traffic intersection four-phase vehicle flow detection method | |
| CN102592138B (en) | Object tracking method for intensive scene based on multi-module sparse projection | |
| CN100435160C (en) | Video image processing method and system for real-time sampling of traffic information | |
| Ullah et al. | Multi-feature-based crowd video modeling for visual event detection | |
| CN104217417B (en) | A kind of method and device of video multi-target tracking | |
| CN105261034A (en) | Method and device for calculating traffic flow on highway | |
| CN106203513A (en) | A kind of based on pedestrian's head and shoulder multi-target detection and the statistical method of tracking | |
| CN103646257A (en) | Video monitoring image-based pedestrian detecting and counting method | |
| CN107169994A (en) | Correlation filtering tracking based on multi-feature fusion | |
| CN102722889B (en) | A kind of image background acquisition methods and device | |
| CN104243916A (en) | Moving object detecting and tracking method based on compressive sensing | |
| Yaghoobi Ershadi et al. | Vehicle tracking and counting system in dusty weather with vibrating camera conditions | |
| CN116434159A (en) | Traffic flow statistics method based on improved YOLO V7 and Deep-Sort | |
| Yang et al. | A fast vehicle counting and traffic volume estimation method based on convolutional neural network | |
| CN112818935B (en) | Multi-lane congestion detection and duration prediction method and system based on deep learning | |
| Li et al. | Intelligent transportation video tracking technology based on computer and image processing technology | |
| Yu et al. | LCNet: a light-weight network for object counting | |
| CN106127800A (en) | Real-time many object tracking methods based on AER imageing sensor and device | |
| CN105023271A (en) | Behavior collective degree measurement method based on multi-clue information | |
| CN110021174A (en) | A kind of vehicle flowrate calculation method for being applicable in more scenes based on video image | |
| CN114694078A (en) | Traffic behavior judgment method based on multi-target tracking | |
| CN103996187A (en) | To-ground moving target photoelectric detection system, and data processing method and image processing method thereof | |
| Liu et al. | Research on vehicle target detection technology based on UAV aerial images |
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 |