CN105528891A - Traffic flow density detection method and system based on unmanned aerial vehicle monitoring - Google Patents
Traffic flow density detection method and system based on unmanned aerial vehicle monitoring Download PDFInfo
- Publication number
- CN105528891A CN105528891A CN201610022176.6A CN201610022176A CN105528891A CN 105528891 A CN105528891 A CN 105528891A CN 201610022176 A CN201610022176 A CN 201610022176A CN 105528891 A CN105528891 A CN 105528891A
- Authority
- CN
- China
- Prior art keywords
- traffic flow
- traffic
- virtual region
- flow density
- video image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G08G1/0112—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
Landscapes
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a traffic flow density detection method based on unmanned aerial vehicle monitoring. The method comprises the following steps: obtaining traffic video image information shot by an unmanned aerial vehicle in a set detection period and a virtual region; carrying out motion segmentation processing on the traffic video image information and summarizing the number of vehicles entering the virtual region in multiple frames of sample images corresponding to multiple moments in selected the traffic video image information and carrying out calculation to obtain the average number of the vehicles entering the virtual region in the detection period; and calculating the traffic flow density in the current detection period according to the corresponding road length and the average number of the vehicles in the virtual region. The invention also discloses a traffic flow density detection system based on unmanned aerial vehicle monitoring. According to the method and system, the real-time traffic flow density of a monitored target road can be calculated accurately, quickly and conveniently, and furthermore, it is convenient to carry out coordinated management on traffic conditions.
Description
Technical field
The present invention relates to technical field of intelligent traffic, particularly relate to the traffic flow Density Detection method and system based on monitoring unmanned.
Background technology
Density is traffic flow very one of important parameter, is also the important indicator differentiating traffic flow modes.Directly can be judged the degree of crowding of traffic by traffic flow density, thus determine to adopt which kind of traffic administration and control measure.Adopt the road occupancy of the vehicle comparatively easily measured indirectly to characterize traffic flow density in actual applications toward contact, vehicle occupancy is higher, and traffic flow density is larger.Wherein, vehicle occupancy specifically comprises space occupancy rate and time occupancy.
Space occupancy rate is generally investigated by eminence Photographic technique, but applicability is not high, and such as eminence Photographic technique can not the height of unrestrictedly absolute altitude video camera, and the link length of also namely taking is restricted; And time occupancy detects mainly through the fixed point such as inductive coil detecting device and obtains, it is the main way investigating traffic flow density at present, time occupancy is larger, then show that traffic flow density is larger, but the method computation process more complicated, and the relation between time occupancy and traffic flow density can not be described quantitatively; Meanwhile, when the magnitude of traffic flow and interval average speed occur comparatively big error time, comparatively big error will be there is in the testing result obtained by this method.
Summary of the invention
Fundamental purpose of the present invention is to provide a kind of traffic flow Density Detection method and system based on monitoring unmanned, and the mode adaptability being intended to solve existing calculating traffic flow density is not high, calculation of complex and the larger technical matters of the error of calculation.
For achieving the above object, the invention provides a kind of traffic flow Density Detection method based on monitoring unmanned, the described traffic flow Density Detection method based on monitoring unmanned comprises:
Obtain the traffic video image information that unmanned plane is taken in the sense cycle and virtual region of setting;
Motion segmentation process is carried out to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
Road section length corresponding to described virtual region and described average traffic number, calculate the traffic flow density in current described sense cycle.
Preferably, comprise before the traffic video image information that described acquisition unmanned plane is taken in the sense cycle and virtual region of setting:
According to the architectural feature of the target road that described unmanned plane is monitored, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
Preferably, described motion segmentation process is carried out to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment and comprise:
Adopt background subtraction and based on dynamic segmentation threshold, motion segmentation process carried out to described traffic video image information, obtaining the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
Adopt multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enter the vehicle fleet size of described virtual region;
According to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
Preferably, the road section length corresponding to described virtual region and described average traffic number build traffic flow density mathematical model, and wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.
Preferably, the traffic flow density mathematical model of described road section length corresponding to described virtual region, described average traffic number and setting, also comprises after calculating the traffic flow density in current described sense cycle:
According to the described traffic flow density calculated and default traffic flow modes threshold value, identify the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
Further, for achieving the above object, the present invention also provides a kind of traffic flow density sensing system based on monitoring unmanned, and the described traffic flow density sensing system based on monitoring unmanned comprises: unmanned plane, traffic flow device for detecting density;
Wherein, described traffic flow device for detecting density comprises:
Traffic image acquisition module, for obtaining the traffic video image information that described unmanned plane is taken in the sense cycle and virtual region of setting;
Traffic image processing module, for carrying out motion segmentation process to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
Traffic flow Density Calculation Module, for the road section length corresponding to described virtual region and described average traffic number, calculates the traffic flow density in current described sense cycle.
Preferably, described traffic flow device for detecting density also comprises:
Virtual region delimit module, for the architectural feature of target road monitored according to described unmanned plane, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
Preferably, described traffic image processing module comprises:
Image segmentation unit, for adopting background subtraction and carrying out motion segmentation process based on dynamic segmentation threshold to described traffic video image information, obtains the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
Vehicle tracking statistic unit, for adopting multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enters the vehicle fleet size of described virtual region;
Average traffic number computing unit, for according to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
Preferably, the road section length corresponding to described virtual region and described average traffic number build traffic flow density mathematical model, and wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.
Preferably, described traffic flow device for detecting density also comprises:
Traffic flow modes identification module, for according to the described traffic flow density calculated and default traffic flow modes threshold value, identifies the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
The present invention adopts unmanned plane to take traffic video image, not only reduce the cost of capture apparatus, can realize detecting the real-time monitoring in emphasis section and the circuit that carries out multiple region by unmanned plane simultaneously, and height, the angle and shooting area etc. of shooting can be adjusted further according to actual needs, thus real-time and the accuracy of traffic data information can be ensured, and then can the accuracy of corresponding raising traffic flow density calculation.In addition, in the present invention, only add up the vehicle fleet size of the virtual region entering setting in sense cycle, thus defined the pixel region scope of image procossing by virtual region, and then can simplify the treatment effeciency of image and the precision that improve vehicle identification.Further, the average traffic number only needing the road section length corresponding to virtual region in the present invention and enter this virtual region can calculate the traffic flow density in the current detection cycle, also namely the direct definition according to traffic flow density calculates traffic flow density, thus simplify traffic flow density calculation process, also ensure the accuracy of result of calculation simultaneously.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of traffic flow Density Detection method first embodiment that the present invention is based on monitoring unmanned;
Fig. 2 for the present invention is based on monitoring unmanned traffic flow Density Detection method one embodiment in average error variation tendency schematic diagram corresponding to different sense cycle;
Fig. 3 for the present invention is based on monitoring unmanned another embodiment of traffic flow Density Detection method in maximum error variation tendency schematic diagram corresponding to different sense cycle;
Fig. 4 is the schematic flow sheet of traffic flow Density Detection method second embodiment that the present invention is based on monitoring unmanned;
Fig. 5 is the schematic diagram of virtual region in traffic flow Density Detection method one embodiment that the present invention is based on monitoring unmanned;
Fig. 6 is the refinement schematic flow sheet of step S20 in Fig. 1;
Fig. 7 is pixel (x, y) and eight neighborhood C8 schematic diagram thereof in motion segmentation process one embodiment in the traffic flow Density Detection method that the present invention is based on monitoring unmanned;
Fig. 8 is the image block schematic diagram calculating illumination factor in the traffic flow Density Detection method that the present invention is based on monitoring unmanned in motion segmentation process one embodiment;
Fig. 9 is the schematic flow sheet of traffic flow Density Detection method the 3rd embodiment that the present invention is based on monitoring unmanned;
Figure 10 is the high-level schematic functional block diagram of traffic flow density sensing system one embodiment that the present invention is based on monitoring unmanned;
Figure 11 is the refinement high-level schematic functional block diagram of traffic flow device for detecting density first embodiment in Figure 10;
Figure 12 is the refinement high-level schematic functional block diagram of traffic flow device for detecting density second embodiment in Figure 10;
Figure 13 is the refinement high-level schematic functional block diagram of traffic image processing module in Figure 11;
Figure 14 is the refinement high-level schematic functional block diagram of traffic flow device for detecting density the 3rd embodiment in Figure 10.
The realization of the object of the invention, functional characteristics and advantage will in conjunction with the embodiments, are described further with reference to accompanying drawing.
Embodiment
Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
With reference to the schematic flow sheet that Fig. 1, Fig. 1 are traffic flow Density Detection method first embodiment that the present invention is based on monitoring unmanned.In the present embodiment, the described traffic flow Density Detection method based on monitoring unmanned comprises:
Step S10, obtains the traffic video image information that unmanned plane is taken in the sense cycle and virtual region of setting;
Traffic flow density specifically refers to a certain moment, the vehicle number on the section of certain unit length in a track or several tracks, specifically for vehicle on reflection road dense degree and weigh road and to get on the bus smooth understanding and considerate condition.According to above definition, traffic flow density is instantaneous value measured on one section of road, and it not only changes over time, also changes with mensuration length of an interval degree.Therefore, preset corresponding sense cycle in the present embodiment, and preferably using the traffic flow density of the mean value of the traffic flow density corresponding to multiple instantaneous moments measured in this sense cycle as this sense cycle.Therefore, traffic flow density corresponding in different sense cycle may be identical, also may not be identical.In addition, virtual region specifically refers to the section for calculating vehicle quantity corresponding in the shooting visual field, and this section (also namely the setting such as length, number of track-lines in section is not limit) corresponding to virtual region is specifically arranged according to actual needs.
In addition, it should be noted that, in the present embodiment, the setting for unmanned plane is not limit, the quantity of such as camera and position, flying height, shooting mode (hovering shooting) etc.In the present embodiment, dock with urban signal controlling system and system for traffic guiding for ease of the traffic video image information taken by unmanned plane, therefore, unmanned plane traffic video image information of taking in the sense cycle and virtual region of setting is synchronously obtained preferably by wireless transmission method.Such as, when current detection end cycle, taken traffic video image information is sent to traffic monitoring platform by mobile communication signal by unmanned plane, and completed vehicle number quantitative statistics in the current detection cycle by traffic monitoring platform, thus obtain the traffic flow density corresponding to the current detection cycle.
In addition, need to further illustrate, for urban road, particularly link length is not the section grown very much, controls, often occur the upper and lower big ups and downs of traffic flow density in the short time owing to being subject to signal, namely during red light, traffic flow density is comparatively large, and green time traffic flow density reduces gradually.Therefore, in order to weaken the detection interference that signal controls to bring, require in the present embodiment that stagger red, the green light of signal of traffic flow data sense cycle opens the bright time, and by the map analysis of metrical error Long-term change trend to determine rational assay intervals (also i.e. sense cycle).In the present embodiment, the temporal sequence of preferred 30s, 1min, 1.5min, 2min, 2.5min, 3min, 3.5min, 4min, 4.5min, 5min, 6min, 7min, 8min, 9min, 10min, 15min is analyzed, and analysis result as Figure 2-3.
According to the variation tendency of Fig. 2-3, when assay intervals is greater than 5min, average error and the maximum error of prediction all tend towards stability, therefore assay intervals needs to be greater than 5min, but assay intervals is oversize, the ageing reduction of traffic flow Density Detection can be made and the value forfeiture detected, because the detection of traffic flow density is often with urban signal controlling system, traffic flow guidance system carries out cooperative cooperating, be whistle control system by the Accurate Prediction of traffic flow density, traffic flow guidance system provides decision-making, if assay intervals is oversize, then cannot provide decision-making in time, and then the delay and ineffective systems that vehicle is larger may be caused.Therefore, comprehensive above-mentioned each side factor, in the present embodiment, preferred 5min is as sense cycle.
Step S20, motion segmentation process is carried out to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
In the present embodiment, for ease of calculating the traffic flow density corresponding to the current detection cycle, therefore, need to process to realize the vehicle number quantitative statistics to entering in this traffic video image information in virtual region to the traffic video image information got.The mode of motion segmentation is preferably adopted to process traffic video image information in the present embodiment, also namely motion segmentation is passed through, the part (background) of the part of moving in video image (vehicle) with geo-stationary is split, thus the vehicle be convenient to entering in virtual region is added up.
In addition, in the present embodiment, for obtaining the average traffic number entering the virtual region of setting in the current detection cycle, therefore, need selected multiple moment and add up in the multiframe sample image corresponding to the plurality of moment vehicle fleet size entered in virtual region respectively, and then obtaining the average traffic number in current detection cycle.
Step S30, the road section length corresponding to described virtual region and described average traffic number, calculate the traffic flow density in current described sense cycle.
In the present embodiment, the average traffic number obtained according to above-mentioned calculating and the road section length corresponding to corresponding virtual region, calculate the traffic flow density in current described sense cycle.Wherein, concrete account form and process are not limit.Such as standardization etc. is carried out to the number of track-lines, lane length etc. in the section corresponding to virtual region.
Based on being generally all carry out indirect judgement traffic flow density by time occupancy or space occupancy in prior art, or calculate traffic flow density by flow and speed formula, but such mode accurately can not reflect current traffic flow density, computation process also compares very complicated simultaneously.
Therefore further alternative, in one embodiment, traffic flow density mathematical model is set up according to the definition of traffic flow density, also the road section length namely corresponding to virtual region and the average traffic number calculated build traffic flow density mathematical model, wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.The instantaneous average traffic number only needing the road section length corresponding to virtual region in this preferred embodiment and enter this virtual region can calculate the traffic flow density in the current detection cycle, also namely the direct definition according to traffic flow density calculates traffic flow density, thus simplify traffic flow density calculation process, also ensure the accuracy of result of calculation simultaneously.What needs further illustrated is, instantaneous average traffic number in virtual region section (also i.e. Q) specifically to refer in the sample image for ease of measuring corresponding to average traffic number and multiple discrete instantaneous moment within the current detection cycle add up the average traffic number entered in virtual region section obtained, also can think that this Q value is and enter the average traffic number of virtual region in the current detection cycle.
In the present embodiment, by the wagon flow situation of unmanned plane (with video camera) Real-time Obtaining road (also i.e. traffic video image information) to catch all vehicles on road in certain area, obtain the vehicle number of moment, and then corresponding traffic flow density can be calculated.Compare other detection methods, unmanned plane relative low price, simultaneously monitoring unmanned detection method can realize the real-time monitoring in emphasis section and can carry out multiple regions circuit and to detect and degree of accuracy is higher.In addition, adopt unmanned plane shooting traffic video image, not only reduce the cost of capture apparatus, the height of shooting, angle and shooting area etc. can be adjusted further according to actual needs simultaneously, thus real-time and the accuracy of traffic data information can be ensured, and then can the accuracy of corresponding raising traffic flow density calculation, such as vehicle supervision department provides decision-making foundation accurately, by coordinating with urban signal controlling system and system for traffic guiding, and then the traffic congestion situation of release portion social connections section.In addition, in the present embodiment, only add up the vehicle fleet size of the virtual region entering setting in sense cycle, thus defined the pixel region scope of image procossing by virtual region, and then can simplify the treatment effeciency of image and the precision that improve vehicle identification.
With reference to the schematic flow sheet that Fig. 4, Fig. 4 are traffic flow Density Detection method second embodiment that the present invention is based on monitoring unmanned.Based on above-described embodiment, in the present embodiment, comprised before above-mentioned steps S10:
Step S01, according to the architectural feature of the target road that described unmanned plane is monitored, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
In the present embodiment, the architectural feature of target road specifically comprises the number of track-lines of road, wagon flow direction, monitor the link length in the visual field and whether have bend etc., the architectural feature of target road both can automatically be identified by system and complete the setting of virtual region, or the setting of virtual region is completed by user's remote manual of manipulation unmanned plane, thus virtual region delimited in the monitoring visual field of unmanned plane, system (such as traffic monitoring platform) then only can be carried out vehicle seizure and carry out vehicle number statistics in virtual region, thus the interference of the outer mobile factor of other target road can be avoided.Virtual region needs to delimit according to the architectural feature of target road, and needs the road section length calibrating corresponding target road, so that subsequent calculations.Wherein, virtual region comprises multiple virtual coil further and intends all in the same way tracks of the corresponding coverage goal road of coil difference, the schematic diagram of virtual region one embodiment as shown in Figure 5.
In the present embodiment, according to road structure, in monitoring visual field, delimit virtual region, and calibrate the physical length in this virtual region section.The road section length of virtual region once demarcate out, and gets the average traffic number in virtual region, can calculate the traffic flow density of this tract section, thus can obtain the traffic flow density in monitored any section in mode more easily.
With reference to the refinement schematic flow sheet that Fig. 6, Fig. 6 are step S20 in Fig. 1.Based on above-described embodiment, in the present embodiment, above-mentioned steps S20 comprises:
Step S201, adopts background subtraction and carries out motion segmentation process based on dynamic segmentation threshold to described traffic video image information, obtaining the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
In the present embodiment, the object of motion segmentation is the part of the part of moving in video image with geo-stationary to split.Because the video camera of unmanned plane and position, road surface keep geo-stationary, background subtraction therefore can be adopted to carry out motion segmentation.It realizes principle substantially: first the stationary part in video image is extracted, as reference background, and then present frame sample image and reference background subtracting, and difference is carried out binary conversion treatment, finally add up the quantity of the pixel changed, and just can think have vehicle to have passed through virtual region after change exceedes noise threshold (also namely fixing segmentation threshold).
In addition, in the present embodiment, the consideration based on following 3: the first, lighting condition there is instability, if strong sunlight is suddenly by cloud, simply subtracts each other and can think that almost entire image all there occurs change.The second, the continuity in motor point, because motion is the object that automobile etc. is larger, so motor point should spatially be connected with other points, those isolated points are likely and are caused by noise.3rd, there is noise in Video Capture transmission channel, even if illumination condition and scenery all do not change, the RGB brightness of pixel still certain change likely occurs.For adapting to the change caused by above-mentioned factor, therefore, based on the consideration of above 3, in the present embodiment, adopting threshold to carry out motion segmentation and specifically carrying out (calculating according to RGB tri-kinds of colors) according to following formula:
As long as above three formulas have an establishment, coordinate points (x, y) just can be judged as the object of motion, otherwise, be exactly static background dot.Understanding for these three formulas is as follows:
1), I
r, t(x, y) denotation coordination is the gray-scale value of red channel in t of the pixel of (x, y), B
r, t(x, y) represents that the pixel of (x, y) is at the gray-scale value of t with reference to background, the integer that wherein, I, B get usually between [0,255], I
g, t(x, y), I
b, t(x, y), B
g, t(x, y), B
b, t(x, y) and I
r, t(x, y), B
r, t(x, y) is similar.
2), α, β are constants.
3), n
t(x, y) is a connectivity factor, the eight neighborhood of its value and point (x, y) (eight namely adjacent with it points, movement properties as shown in Figure 7) is relevant,
wherein M (x ', y ') is a binary.If fruit dot (x ', y ') is motor point, so M (x ', y ') is 1; Otherwise be 0.Because the relation of scanning sequency, the value coming (x, y) four points below can be determined by the movement properties of previous frame.Be not difficult to know, if pending point (x, y) peripheral motor point is more, the partition value of (x, y) is less, and (x, y) is more easily judged to motor point; Vice versa.
4), σ
t(x, y) is standard deviation, and its definition is:
Wherein, μ (x, y) is mathematical expectation.Sometimes in order to save memory headroom and operation time, standard deviation can utilize the method increased progressively to carry out calculating (for red channel):
σ
2 R,t(x,y)＝(1-ρ)σ
2 R,t-1(x,y)+ρ[I
R,t(x,y)-B
R,t(x,y)]
2
Wherein ρ can be the variable of a time, also can be similar to and think a constant.The calculating of the standard deviation of other two Color Channels is identical with red channel.
5), k
t(x, y) is the consideration of collating condition change.As shown in Figure 8, first the image when former frame is divided into square area W little one by one
i,j, the length of side of each window is m pixel, and the value of m is determined according to the size of object to be detected, for the detection of vehicle, generally gets 5-9.(x, y) must belong to one in all square area, is designated as S, so
Step S202, adopts multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enters the vehicle fleet size of described virtual region;
By after above-mentioned motion segmentation process obtain in multiframe sample image, entering the pixel set corresponding to the vehicle in virtual region respectively, and then can vehicle number quantitative statistics be carried out, but due to multi-section vehicle may be there is in virtual region, and catch vehicle only by the only virtual coil in each track and cannot meet many vehicles and catch demand simultaneously, therefore, multiple target tracking algorithm is preferably adopted to catch vehicles all in virtual region and count in the present embodiment.Multiple target tracking algorithm is same as the prior art, does not therefore do too much repeating.
Step S203, according to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
In the present embodiment, based on the traffic flow density mathematical model in above-described embodiment, therefore, need to obtain the average traffic number entering described virtual region in described sense cycle, obtain especially by the average traffic quantity entered in this virtual region in the every frame sample image of calculating.Also in sense cycle, namely obtain multiple moment enters vehicle number in virtual region, then asks for arithmetic average.The vehicle number entered in virtual region that such as CCTV camera captures for 5 times in sense cycle 5min is respectively 18,32,23,37,20, the average traffic number entered in this sense cycle 5min in virtual region then reflected is 24, is also that the Q value in above-mentioned traffic flow density mathematical model is 24.
In the present embodiment, by delimiting virtual region in video, and multiple target tracking algorithm only carries out multiple target tracking seizure to the multiple vehicles in region and carries out quantity statistics, thus can the degree of accuracy that catches of corresponding raising vehicle.In addition, by carrying out motion segmentation to traffic video image information, and add up the vehicle fleet size entered in virtual region, can according to traffic flow density mathematical model of the present invention, calculate the traffic flow density in the current detection cycle, thus simplify the determination mode of traffic flow density, also make the accuracy of traffic flow density get a promotion simultaneously.
With reference to the schematic flow sheet that Fig. 9, Fig. 9 are traffic flow Density Detection method the 3rd embodiment that the present invention is based on monitoring unmanned.Based on above-described embodiment, in the present embodiment, also comprise after above-mentioned steps S30:
Step S40, according to the described traffic flow density calculated and default traffic flow modes threshold value, identifies the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
In the present embodiment, after getting the traffic flow density of target road, target road traffic flow modes can be differentiated according to traffic behavior grade scale.Because road traffic current density is subject to the impact of the factor such as road structure and road vehicle composition, the traffic flow density critical value of different road settings and the variation characteristic of traffic flow density should be different.So determine that the traffic flow density critical value of certain specific road section is very necessary according to the traffic flow density data of actual measurement.And in practice, after the general traffic flow density according to recording compares with the traffic flow density critical value of setting, directly judge the traffic flow conditions of road.As shown in table 1 below be a certain bar road traffic flow modes judge critical value, wherein 0,12,44,90 is traffic flow modes threshold value, by comparing the traffic flow modes threshold value of traffic flow density and the setting calculated, current traffic flow modes can be determined, such as, if the Current traffic current density calculated is 55, then the traffic flow modes that can be defined as the target road of current monitoring is slight crowding.
Table 1
| Traffic flow modes | Smooth and easy | Jogging | Slight crowding | Heavy congestion |
| Traffic flow density (/ track * km) | 0－12 | 12－44 | 44－90 | ＞90 |
In the present embodiment, by the wagon flow situation of unmanned plane Real-time Obtaining road, and adopt all vehicles in multiple goal capturing technology statistics target road in certain virtual region, and the vehicle number obtaining multiple time instant is to obtain the average traffic number in sense cycle, and then calculate traffic flow density, directly provide current traffic flow modes further so that user reads and makes subjective judgement simultaneously.Monitoring unmanned detection method can be monitored in real time to emphasis section, crossing, the circuit that simultaneously also can carry out multiple tract section detects, maneuverability is very strong, monitor wide, and the real-time traffic condition in section passback Surveillance center can be thought that vehicle supervision department provides decision-making foundation accurately.
With reference to the high-level schematic functional block diagram that Figure 10, Figure 10 are traffic flow density sensing system one embodiment that the present invention is based on monitoring unmanned.In the present embodiment, the described traffic flow density sensing system based on monitoring unmanned comprises: unmanned plane 10, traffic flow device for detecting density 20; Wherein, unmanned plane 10 is not limit with the setting of traffic flow device for detecting density 20, such as, traffic flow device for detecting density 20 is set directly on unmanned plane 10, or unmanned plane 10 is connected to carry out exchanges data by mobile communication signal with traffic flow device for detecting density 20.
As shown in figure 11, wherein, described traffic flow device for detecting density 20 comprises:
Traffic image acquisition module 201, for obtaining the traffic video image information that described unmanned plane is taken in the sense cycle and virtual region of setting;
Traffic flow density specifically refers to a certain moment, the vehicle number on the section of certain unit length in a track or several tracks, specifically for vehicle on reflection road dense degree and weigh road and to get on the bus smooth understanding and considerate condition.According to above definition, traffic flow density is instantaneous value measured on one section of road, and it not only changes over time, also changes with mensuration length of an interval degree.Therefore, preset corresponding sense cycle in the present embodiment, and preferably using the traffic flow density of the mean value of the traffic flow density corresponding to multiple instantaneous moments measured in this sense cycle as this sense cycle.Therefore, traffic flow density corresponding in different sense cycle may be identical, also may not be identical.In addition, virtual region specifically refers to the section for calculating vehicle quantity corresponding in the shooting visual field, and this section (also namely the setting such as length, number of track-lines in section is not limit) corresponding to virtual region is specifically arranged according to actual needs.
In addition, it should be noted that, in the present embodiment, the setting for unmanned plane is not limit, the quantity of such as camera and position, flying height, shooting mode (hovering shooting) etc.In the present embodiment, dock with urban signal controlling system and system for traffic guiding for ease of the traffic video image information taken by unmanned plane, therefore, unmanned plane traffic video image information of taking in the sense cycle and virtual region of setting is synchronously obtained preferably by wireless transmission method.Such as, when current detection end cycle, taken traffic video image information is sent to traffic monitoring platform by mobile communication signal by unmanned plane, and completed vehicle number quantitative statistics in the current detection cycle by traffic monitoring platform, thus obtain the traffic flow density corresponding to the current detection cycle.
In addition, need to further illustrate, for urban road, particularly link length is not the section grown very much, controls, often occur the upper and lower big ups and downs of traffic flow density in the short time owing to being subject to signal, namely during red light, traffic flow density is comparatively large, and green time traffic flow density reduces gradually.Therefore, in order to weaken the detection interference that signal controls to bring, require in the present embodiment that stagger red, the green light of signal of traffic flow data sense cycle opens the bright time, and by the map analysis of metrical error Long-term change trend to determine rational assay intervals (also i.e. sense cycle).In the present embodiment, the temporal sequence of preferred 30s, 1min, 1.5min, 2min, 2.5min, 3min, 3.5min, 4min, 4.5min, 5min, 6min, 7min, 8min, 9min, 10min, 15min is analyzed, and analysis result as Figure 2-3.
According to the variation tendency of Fig. 2-3, when assay intervals is greater than 5min, average error and the maximum error of prediction all tend towards stability, therefore assay intervals needs to be greater than 5min, but assay intervals is oversize, the ageing reduction of traffic flow Density Detection can be made and the value forfeiture detected, because the detection of traffic flow density is often with urban signal controlling system, traffic flow guidance system carries out cooperative cooperating, be whistle control system by the Accurate Prediction of traffic flow density, traffic flow guidance system provides decision-making, if assay intervals is oversize, then cannot provide decision-making in time, and then the delay and ineffective systems that vehicle is larger may be caused.Therefore, comprehensive above-mentioned each side factor, in the present embodiment, preferred 5min is as sense cycle.
Traffic image processing module 202, for carrying out motion segmentation process to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
In the present embodiment, for ease of calculating the traffic flow density corresponding to the current detection cycle, therefore, traffic image processing module 202 processes to realize the vehicle number quantitative statistics to entering in this traffic video image information in virtual region to the traffic video image information got.The mode of motion segmentation is preferably adopted to process traffic video image information in the present embodiment, also namely motion segmentation is passed through, the part (background) of the part of moving in video image (vehicle) with geo-stationary is split, thus the vehicle be convenient to entering in virtual region is added up.
In addition, in the present embodiment, for obtaining the average traffic number entering the virtual region of setting in the current detection cycle, therefore, need selected multiple moment and add up in the multiframe sample image corresponding to the plurality of moment vehicle fleet size entered in virtual region respectively, and then obtaining the average traffic number in current detection cycle.
Traffic flow Density Calculation Module 203, for the road section length corresponding to described virtual region and described average traffic number, calculates the traffic flow density in current described sense cycle.
In the present embodiment, the average traffic number that traffic flow Density Calculation Module 203 obtains according to above-mentioned calculating and the road section length corresponding to corresponding virtual region, calculate the traffic flow density in current described sense cycle.Wherein, concrete account form and process are not limit.Such as standardization etc. is carried out to the number of track-lines, lane length etc. in the section corresponding to virtual region.
Based on being generally all carry out indirect judgement traffic flow density by time occupancy or space occupancy in prior art, or calculate traffic flow density by flow and speed formula, but such mode accurately can not reflect current traffic flow density, computation process also compares very complicated simultaneously.
Therefore further alternative, in traffic flow density sensing system one embodiment that the present invention is based on monitoring unmanned, traffic flow density mathematical model is set up according to the definition of traffic flow density, also namely the road section length of traffic flow Density Calculation Module 203 corresponding to virtual region builds traffic flow density mathematical model with the average traffic number calculated, wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.The instantaneous average traffic number only needing the road section length corresponding to virtual region in this preferred embodiment and enter this virtual region can calculate the traffic flow density in the current detection cycle, also namely the direct definition according to traffic flow density calculates traffic flow density, thus simplify traffic flow density calculation process, also ensure the accuracy of result of calculation simultaneously.What needs further illustrated is, instantaneous average traffic number in virtual region section (also i.e. Q) specifically to refer in the sample image for ease of measuring corresponding to average traffic number and multiple discrete instantaneous moment within the current detection cycle add up the average traffic number entered in virtual region section obtained, also can think that this Q value is and enter the average traffic number of virtual region in the current detection cycle.
In the present embodiment, by the wagon flow situation of unmanned plane (with video camera) Real-time Obtaining road (also i.e. traffic video image information) to catch all vehicles on road in certain area, obtain the vehicle number of moment, and then corresponding traffic flow density can be calculated.Compare other detection methods, unmanned plane relative low price, simultaneously monitoring unmanned detection method can realize the real-time monitoring in emphasis section and can carry out multiple regions circuit and to detect and degree of accuracy is higher.In addition, adopt unmanned plane shooting traffic video image, not only reduce the cost of capture apparatus, the height of shooting, angle and shooting area etc. can be adjusted further according to actual needs simultaneously, thus real-time and the accuracy of traffic data information can be ensured, and then can the accuracy of corresponding raising traffic flow density calculation, such as vehicle supervision department provides decision-making foundation accurately, by coordinating with urban signal controlling system and system for traffic guiding, and then the traffic congestion situation of release portion social connections section.In addition, in the present embodiment, only add up the vehicle fleet size of the virtual region entering setting in sense cycle, thus defined the pixel region scope of image procossing by virtual region, and then can simplify the treatment effeciency of image and the precision that improve vehicle identification.
With reference to the refinement high-level schematic functional block diagram that Figure 12, Figure 12 are traffic flow device for detecting density second embodiment in Figure 10.In the present embodiment, described traffic flow device for detecting density also comprises:
Virtual region delimit module 204, for the architectural feature of target road monitored according to described unmanned plane, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
In the present embodiment, the architectural feature of target road specifically comprises the link length in the number of track-lines of road, wagon flow direction, the monitoring visual field and whether has bend etc., the architectural feature of target road delimited module 204 by virtual region and automatically identified and the setting completing virtual region, thus virtual region delimited in the monitoring visual field of unmanned plane, system (such as traffic monitoring platform) then only can be carried out vehicle seizure and carry out vehicle number statistics in virtual region, thus can avoid the interference of the outer mobile factor of other target road.Virtual region needs to delimit according to the architectural feature of target road, and needs the road section length calibrating corresponding target road, so that subsequent calculations.Wherein, virtual region comprises multiple virtual coil further and intends all in the same way tracks of the corresponding coverage goal road of coil difference, the schematic diagram of virtual region one embodiment as shown in Figure 5.
In the present embodiment, according to road structure, in monitoring visual field, delimit virtual region, and calibrate the physical length in this virtual region section.The road section length of virtual region once demarcate out, and gets the average traffic number in virtual region, can calculate the traffic flow density of this tract section, thus can obtain the traffic flow density in monitored any section in mode more easily.
With reference to the refinement high-level schematic functional block diagram that Figure 13, Figure 13 are traffic image processing module in Figure 11.Based on above-described embodiment, in the present embodiment, described traffic image processing module 202 comprises:
Image segmentation unit 2021, for adopting background subtraction and carrying out motion segmentation process based on dynamic segmentation threshold to described traffic video image information, obtain the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
In the present embodiment, the object of motion segmentation is the part of the part of moving in video image with geo-stationary to split.Because the video camera of unmanned plane and position, road surface keep geo-stationary, background subtraction therefore can be adopted to carry out motion segmentation.It realizes principle substantially: first the stationary part in video image is extracted, as reference background, and then present frame sample image and reference background subtracting, and difference is carried out binary conversion treatment, finally add up the quantity of the pixel changed, and just can think have vehicle to have passed through virtual region after change exceedes noise threshold (also namely fixing segmentation threshold).
In addition, in the present embodiment, the consideration based on following 3: the first, lighting condition there is instability, if strong sunlight is suddenly by cloud, simply subtracts each other and can think that almost entire image all there occurs change.The second, the continuity in motor point, because motion is the object that automobile etc. is larger, so motor point should spatially be connected with other points, those isolated points are likely and are caused by noise.3rd, there is noise in Video Capture transmission channel, even if illumination condition and scenery all do not change, the RGB brightness of pixel still certain change likely occurs.For adapting to the change caused by above-mentioned factor, therefore, based on the consideration of above 3, in the present embodiment, adopting threshold to carry out motion segmentation and specifically carrying out (calculating according to RGB tri-kinds of colors) according to following formula:
As long as above three formulas have an establishment, coordinate points (x, y) just can be judged as the object of motion, otherwise, be exactly static background dot.Understanding for these three formulas is as follows:
1), I
r, t(x, y) denotation coordination is the gray-scale value of red channel in t of the pixel of (x, y), B
r, t(x, y) represents that the pixel of (x, y) is at the gray-scale value of t with reference to background, the integer that wherein, I, B get usually between [0,255], I
g, t(x, y), I
b, t(x, y), B
g, t(x, y), B
b, t(x, y) and I
r, t(x, y), B
r, t(x, y) is similar.
2), α, β are constants.
3), n
t(x, y) is a connectivity factor, the eight neighborhood of its value and point (x, y) (eight namely adjacent with it points, movement properties as shown in Figure 7) is relevant,
wherein M (x ', y ') is a binary.If fruit dot (x ', y ') is motor point, so M (x ', y ') is 1; Otherwise be 0.Because the relation of scanning sequency, the value coming (x, y) four points below can be determined by the movement properties of previous frame.Be not difficult to know, if pending point (x, y) peripheral motor point is more, the partition value of (x, y) is less, and (x, y) is more easily judged to motor point; Vice versa.
4), σ
t(x, y) is standard deviation, and its definition is:
Wherein, μ (x, y) is mathematical expectation.Sometimes in order to save memory headroom and operation time, standard deviation can utilize the method increased progressively to carry out calculating (for red channel):
σ
2 R,t(x,y)＝(1-ρ)σ
2 R,t-1(x,y)+ρ[I
R,t(x,y)-B
R,t(x,y)]
2
Wherein ρ can be the variable of a time, also can be similar to and think a constant.The calculating of the standard deviation of other two Color Channels is identical with red channel.
5), k
t(x, y) is the consideration of collating condition change.As shown in Figure 8, first the image when former frame is divided into square area W little one by one
i,j, the length of side of each window is m pixel, and the value of m is determined according to the size of object to be detected, for the detection of vehicle, generally gets 5-9.(x, y) must belong to one in all square area, is designated as S, so
Vehicle tracking statistic unit 2022, for adopting multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enters the vehicle fleet size of described virtual region;
By after above-mentioned motion segmentation process obtain in multiframe sample image, entering the pixel set corresponding to the vehicle in virtual region respectively, and then can vehicle number quantitative statistics be carried out, but due to multi-section vehicle may be there is in virtual region, and catch vehicle only by the only virtual coil in each track and cannot meet many vehicles and catch demand simultaneously, therefore, multiple target tracking algorithm is preferably adopted to catch vehicles all in virtual region and count in the present embodiment.Multiple target tracking algorithm is same as the prior art, does not therefore do too much repeating.
Average traffic number computing unit 2023, for according to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
In the present embodiment, based on the traffic flow density mathematical model in above-described embodiment, therefore, need to obtain the average traffic number entering described virtual region in described sense cycle, obtain especially by the average traffic quantity entered in this virtual region in the every frame sample image of calculating.Also in sense cycle, namely obtain multiple moment enters vehicle number in virtual region, then asks for arithmetic average.The vehicle number entered in virtual region that such as CCTV camera captures for 5 times in sense cycle 5min is respectively 18,32,23,37,20, the average traffic number entered in this sense cycle 5min in virtual region then reflected is 24, is also that the Q value in above-mentioned traffic flow density mathematical model is 24.
In the present embodiment, by delimiting virtual region in video, and multiple target tracking algorithm only carries out multiple target tracking seizure to the multiple vehicles in region and carries out quantity statistics, thus can the degree of accuracy that catches of corresponding raising vehicle.In addition, by carrying out motion segmentation to traffic video image information, and add up the vehicle fleet size entered in virtual region, can according to traffic flow density mathematical model of the present invention, calculate the traffic flow density in the current detection cycle, thus simplify the determination mode of traffic flow density, also make the accuracy of traffic flow density get a promotion simultaneously.
With reference to the refinement high-level schematic functional block diagram that Figure 14, Figure 14 are traffic flow device for detecting density the 3rd embodiment in Figure 10.Based on above-described embodiment, in the present embodiment, described traffic flow device for detecting density also comprises:
Traffic flow modes identification module 205, for according to the described traffic flow density calculated and default traffic flow modes threshold value, identifies the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
In the present embodiment, after getting the traffic flow density of target road, target road traffic flow modes can be differentiated according to traffic behavior grade scale.Because road traffic current density is subject to the impact of the factor such as road structure and road vehicle composition, the traffic flow density critical value of different road settings and the variation characteristic of traffic flow density should be different.So determine that the traffic flow density critical value of certain specific road section is very necessary according to the traffic flow density data of actual measurement.And in practice, after the general traffic flow density according to recording compares with the traffic flow density critical value of setting, directly judge the traffic flow conditions of road.As shown in table 2 below be a certain bar road traffic flow modes judge critical value, wherein 0,12,44,90 is traffic flow modes threshold value, by comparing the traffic flow modes threshold value of traffic flow density and the setting calculated, current traffic flow modes can be determined, such as, if the Current traffic current density calculated is 55, then the traffic flow modes that can be defined as the target road of current monitoring is slight crowding.
Table 2
| Traffic flow modes | Smooth and easy | Jogging | Slight crowding | Heavy congestion |
| Traffic flow density (/ track * km) | 0－12 | 12－44 | 44－90 | ＞90 |
In the present embodiment, by the wagon flow situation of unmanned plane Real-time Obtaining road, and adopt all vehicles in multiple goal capturing technology statistics target road in certain virtual region, and the vehicle number obtaining multiple time instant is to obtain the average traffic number in sense cycle, and then calculate traffic flow density, directly provide current traffic flow modes further so that user reads and makes subjective judgement simultaneously.Monitoring unmanned detection method can be monitored in real time to emphasis section, crossing, the circuit that simultaneously also can carry out multiple tract section detects, maneuverability is very strong, monitor wide, and the real-time traffic condition in section passback Surveillance center can be thought that vehicle supervision department provides decision-making foundation accurately.
These are only the preferred embodiments of the present invention; not thereby the scope of the claims of the present invention is limited; every utilize instructions of the present invention and accompanying drawing content to do equivalent structure or equivalent flow process conversion; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.
Claims (10)
1. based on a traffic flow Density Detection method for monitoring unmanned, it is characterized in that, the described traffic flow Density Detection method based on monitoring unmanned comprises:
Obtain the traffic video image information that unmanned plane is taken in the sense cycle and virtual region of setting;
Motion segmentation process is carried out to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
Road section length corresponding to described virtual region and described average traffic number, calculate the traffic flow density in current described sense cycle.
2. as claimed in claim 1 based on the traffic flow Density Detection method of monitoring unmanned, it is characterized in that, comprise before the traffic video image information that described acquisition unmanned plane is taken in the sense cycle and virtual region of setting:
According to the architectural feature of the target road that described unmanned plane is monitored, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
3. as claimed in claim 2 based on the traffic flow Density Detection method of monitoring unmanned, it is characterized in that, described motion segmentation process is carried out to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment and comprise:
Adopt background subtraction and based on dynamic segmentation threshold, motion segmentation process carried out to described traffic video image information, obtaining the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
Adopt multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enter the vehicle fleet size of described virtual region;
According to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
4. the traffic flow Density Detection method based on monitoring unmanned according to any one of claim 1-3, it is characterized in that, road section length corresponding to described virtual region and described average traffic number build traffic flow density mathematical model, and wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.
5. as claimed in claim 4 based on the traffic flow Density Detection method of monitoring unmanned, it is characterized in that, the traffic flow density mathematical model of described road section length corresponding to described virtual region, described average traffic number and setting, also comprises after calculating the traffic flow density in current described sense cycle:
According to the described traffic flow density calculated and default traffic flow modes threshold value, identify the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
6. based on a traffic flow density sensing system for monitoring unmanned, it is characterized in that, the described traffic flow density sensing system based on monitoring unmanned comprises: unmanned plane, traffic flow device for detecting density;
Wherein, described traffic flow device for detecting density comprises:
Traffic image acquisition module, for obtaining the traffic video image information that described unmanned plane is taken in the sense cycle and virtual region of setting;
Traffic image processing module, for carrying out motion segmentation process to described traffic video image information, and statistics enter the vehicle fleet size in described virtual region and calculate the average traffic number entering described virtual region in described sense cycle in multiframe sample image in the described traffic video image information selected corresponding to multiple moment;
Traffic flow Density Calculation Module, for the road section length corresponding to described virtual region and described average traffic number, calculates the traffic flow density in current described sense cycle.
7., as claimed in claim 6 based on the traffic flow density sensing system of monitoring unmanned, it is characterized in that, described traffic flow device for detecting density also comprises:
Virtual region delimit module, for the architectural feature of target road monitored according to described unmanned plane, described virtual region delimited and the road section length of demarcating corresponding to described virtual region in the monitoring visual field of described unmanned plane, wherein, described virtual region comprises multiple virtual coil and the described virtual coil corresponding all in the same way tracks covering described target road respectively.
8., as claimed in claim 7 based on the traffic flow density sensing system of monitoring unmanned, it is characterized in that, described traffic image processing module comprises:
Image segmentation unit, for adopting background subtraction and carrying out motion segmentation process based on dynamic segmentation threshold to described traffic video image information, obtains the pixel set corresponding to vehicle entered respectively in multiframe sample image in described virtual region;
Vehicle tracking statistic unit, for adopting multiple target tracking algorithm to follow the tracks of described pixel set, and corresponding statistics enters the vehicle fleet size of described virtual region;
Average traffic number computing unit, for according to adding up the vehicle fleet size that obtains, calculating in every frame sample image and entering average traffic quantity in described virtual region to obtain entering the average traffic number of described virtual region in described sense cycle.
9. the traffic flow density sensing system based on monitoring unmanned according to any one of claim 6-8, it is characterized in that, road section length corresponding to described virtual region and described average traffic number build traffic flow density mathematical model, and wherein, described traffic flow density mathematical model is:
K＝Q/L；
Wherein, K is traffic flow density, and L is the road section length of virtual region, and Q is the instantaneous average traffic number in virtual region section.
10., as claimed in claim 9 based on the traffic flow density sensing system of monitoring unmanned, it is characterized in that, described traffic flow device for detecting density also comprises:
Traffic flow modes identification module, for according to the described traffic flow density calculated and default traffic flow modes threshold value, identifies the traffic flow modes of the target road that described unmanned plane is monitored in current described sense cycle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610022176.6A CN105528891A (en) | 2016-01-13 | 2016-01-13 | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610022176.6A CN105528891A (en) | 2016-01-13 | 2016-01-13 | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105528891A true CN105528891A (en) | 2016-04-27 |
Family
ID=55771086
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610022176.6A Pending CN105528891A (en) | 2016-01-13 | 2016-01-13 | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105528891A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761494A (en) * | 2016-05-12 | 2016-07-13 | 招商局重庆交通科研设计院有限公司 | Abnormal traffic information collecting method based on unmanned aerial vehicle |
| CN105788269A (en) * | 2016-05-12 | 2016-07-20 | 招商局重庆交通科研设计院有限公司 | Unmanned aerial vehicle-based abnormal traffic identification method |
| CN106097726A (en) * | 2016-08-23 | 2016-11-09 | 苏州科达科技股份有限公司 | The detection determination in region, traffic information detection method and device |
| CN106529500A (en) * | 2016-11-28 | 2017-03-22 | 中控智慧科技股份有限公司 | Information processing method and system |
| CN106875707A (en) * | 2017-04-12 | 2017-06-20 | 武汉理工大学 | A kind of aerial traffic police's system for emergent traffic scene |
| CN107248296A (en) * | 2017-07-13 | 2017-10-13 | 南京航空航天大学 | A kind of video traffic flow statistical method based on unmanned plane and temporal aspect |
| CN107820010A (en) * | 2017-11-17 | 2018-03-20 | 英业达科技有限公司 | Photography counting device |
| CN108320510A (en) * | 2018-04-03 | 2018-07-24 | 深圳市智绘科技有限公司 | One kind being based on unmanned plane video traffic information statistical method and system |
| CN108364466A (en) * | 2018-02-11 | 2018-08-03 | 金陵科技学院 | A kind of statistical method of traffic flow based on unmanned plane traffic video |
| 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 |
| CN110223518A (en) * | 2019-06-21 | 2019-09-10 | 北京航空航天大学 | A kind of intensive section traffic flow guidance method of vehicle towards 3 D stereo traffic system |
| CN111145550A (en) * | 2019-12-27 | 2020-05-12 | 江苏省新通智能交通科技发展有限公司 | Traffic state information acquisition method based on unmanned aerial vehicle |
| CN113506264A (en) * | 2021-07-07 | 2021-10-15 | 北京航空航天大学 | Road vehicle number identification method and device |
| CN113763711A (en) * | 2021-10-13 | 2021-12-07 | 深圳联和智慧科技有限公司 | Unmanned aerial vehicle traffic monitoring method and system based on city management |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510358A (en) * | 2009-03-20 | 2009-08-19 | 吉林大学 | Method and apparatus for processing real time statistical vehicle flowrate using video image |
| CN101587646A (en) * | 2008-05-21 | 2009-11-25 | 上海新联纬讯科技发展有限公司 | Method and system of traffic flow detection based on video identification technology |
| CN102436738A (en) * | 2011-09-26 | 2012-05-02 | 同济大学 | Traffic monitoring device based on unmanned aerial vehicle (UAV) |
| CN103413444A (en) * | 2013-08-26 | 2013-11-27 | 深圳市川大智胜科技发展有限公司 | Traffic flow surveying and handling method based on unmanned aerial vehicle high-definition video |
| GB2508279A (en) * | 2012-10-04 | 2014-05-28 | Honeywell Int Inc | Detecting motion in high resolution video images |
| CN104900054A (en) * | 2015-04-30 | 2015-09-09 | 深圳市中盟科技股份有限公司 | Traffic state determining method and apparatus |
-
2016
- 2016-01-13 CN CN201610022176.6A patent/CN105528891A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101587646A (en) * | 2008-05-21 | 2009-11-25 | 上海新联纬讯科技发展有限公司 | Method and system of traffic flow detection based on video identification technology |
| CN101510358A (en) * | 2009-03-20 | 2009-08-19 | 吉林大学 | Method and apparatus for processing real time statistical vehicle flowrate using video image |
| CN102436738A (en) * | 2011-09-26 | 2012-05-02 | 同济大学 | Traffic monitoring device based on unmanned aerial vehicle (UAV) |
| GB2508279A (en) * | 2012-10-04 | 2014-05-28 | Honeywell Int Inc | Detecting motion in high resolution video images |
| CN103413444A (en) * | 2013-08-26 | 2013-11-27 | 深圳市川大智胜科技发展有限公司 | Traffic flow surveying and handling method based on unmanned aerial vehicle high-definition video |
| CN104900054A (en) * | 2015-04-30 | 2015-09-09 | 深圳市中盟科技股份有限公司 | Traffic state determining method and apparatus |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761494A (en) * | 2016-05-12 | 2016-07-13 | 招商局重庆交通科研设计院有限公司 | Abnormal traffic information collecting method based on unmanned aerial vehicle |
| CN105788269A (en) * | 2016-05-12 | 2016-07-20 | 招商局重庆交通科研设计院有限公司 | Unmanned aerial vehicle-based abnormal traffic identification method |
| CN105761494B (en) * | 2016-05-12 | 2019-01-01 | 招商局重庆交通科研设计院有限公司 | Abnormal traffic information collecting method based on unmanned plane |
| CN106097726A (en) * | 2016-08-23 | 2016-11-09 | 苏州科达科技股份有限公司 | The detection determination in region, traffic information detection method and device |
| CN106529500A (en) * | 2016-11-28 | 2017-03-22 | 中控智慧科技股份有限公司 | Information processing method and system |
| CN106875707A (en) * | 2017-04-12 | 2017-06-20 | 武汉理工大学 | A kind of aerial traffic police's system for emergent traffic scene |
| CN107248296A (en) * | 2017-07-13 | 2017-10-13 | 南京航空航天大学 | A kind of video traffic flow statistical method based on unmanned plane and temporal aspect |
| CN107820010A (en) * | 2017-11-17 | 2018-03-20 | 英业达科技有限公司 | Photography counting device |
| CN108364466A (en) * | 2018-02-11 | 2018-08-03 | 金陵科技学院 | A kind of statistical method of traffic flow based on unmanned plane traffic video |
| 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 |
| CN108320510A (en) * | 2018-04-03 | 2018-07-24 | 深圳市智绘科技有限公司 | One kind being based on unmanned plane video traffic information statistical method and system |
| 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 |
| CN110223518A (en) * | 2019-06-21 | 2019-09-10 | 北京航空航天大学 | A kind of intensive section traffic flow guidance method of vehicle towards 3 D stereo traffic system |
| CN111145550A (en) * | 2019-12-27 | 2020-05-12 | 江苏省新通智能交通科技发展有限公司 | Traffic state information acquisition method based on unmanned aerial vehicle |
| CN113506264A (en) * | 2021-07-07 | 2021-10-15 | 北京航空航天大学 | Road vehicle number identification method and device |
| CN113506264B (en) * | 2021-07-07 | 2023-08-29 | 北京航空航天大学 | Road vehicle number identification method and device |
| CN113763711A (en) * | 2021-10-13 | 2021-12-07 | 深圳联和智慧科技有限公司 | Unmanned aerial vehicle traffic monitoring method and system based on city management |
| CN113763711B (en) * | 2021-10-13 | 2022-06-14 | 深圳联和智慧科技有限公司 | Unmanned aerial vehicle traffic monitoring method and system based on city management |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105528891A (en) | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring | |
| US10386476B2 (en) | Obstacle detection method and apparatus for vehicle-mounted radar system | |
| CN109064495A (en) | A kind of bridge floor vehicle space time information acquisition methods based on Faster R-CNN and video technique | |
| CN102862574B (en) | The method of vehicle active safety is realized based on smart mobile phone | |
| CN108345855B (en) | Lane line pressing detection method and system | |
| US11322028B2 (en) | Radar-augmentation of parking space sensors | |
| CN109426788B (en) | Queuing length detection method and device and server | |
| CN100454355C (en) | Video method for collecting information of vehicle flowrate on road in real time | |
| CN103954542A (en) | PM2.5 (Particulate Matter2.5) concentration detector based on definition evaluation without reference image | |
| CN104900054A (en) | Traffic state determining method and apparatus | |
| CN103593678A (en) | Long-span bridge vehicle dynamic load distribution detection method | |
| CN110255318B (en) | Method for detecting idle articles in elevator car based on image semantic segmentation | |
| CN113192336B (en) | Road congestion condition detection method taking robust vehicle target detection as core | |
| CN107240268A (en) | A kind of detection method of high speed Emergency Vehicle Lane vehicles peccancy towards unmanned plane | |
| CN107818685A (en) | A kind of method that state of motion of vehicle is obtained based on Vehicular video | |
| CN110120155A (en) | A kind of chemical industry plant area vehicle overload overload intelligent monitoring and alarming system | |
| CN111896549A (en) | Building crack monitoring system and method based on machine learning | |
| CN105173945A (en) | Elevator car manned space detection method | |
| CN112528208B (en) | Weighing-free AI intelligent recognition truck overload estimation method, device and system | |
| KR20140144047A (en) | System and method for estimating traffic characteristics information in the road network on a real-time basis | |
| CN117058600B (en) | Regional bridge group traffic load identification method and system | |
| CN109598947A (en) | A kind of vehicle identification method and system | |
| CN113689505B (en) | Method and device for adjusting exposure parameters of vehicle-mounted camera based on positioning information | |
| CN110660229A (en) | Vehicle speed measuring method and device and vehicle | |
| CN110411686B (en) | Bridge static and dynamic image holographic property health monitoring and diagnosis method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| CB02 | Change of applicant information |
Address after: 518000, Shenzhen, Guangdong, Nanshan District Nanshan commercial and cultural center Tianli Central Business Plaza (phase two) C-3005 Applicant after: SHENZHEN UNIHZ TECHNOLOGIES CO., LTD. Address before: 518000, Shenzhen, Guangdong, Nanshan District Nanshan commercial and cultural center Tianli Central Business Plaza (phase two) C-3005 Applicant before: Zhongmeng Science-Technology Co., Ltd., Shenzhen |
|
| COR | Change of bibliographic data | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160427 |