CN107316462A - A kind of flow statistical method and device - Google Patents
A kind of flow statistical method and device Download PDFInfo
- Publication number
- CN107316462A CN107316462A CN201710761698.2A CN201710761698A CN107316462A CN 107316462 A CN107316462 A CN 107316462A CN 201710761698 A CN201710761698 A CN 201710761698A CN 107316462 A CN107316462 A CN 107316462A
- Authority
- CN
- China
- Prior art keywords
- statistics
- objects
- field picture
- frame
- 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
- G08G1/0116—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from roadside infrastructure, e.g. beacons
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a kind of flow statistical method and device, this method includes:Object identifying model of the generation corresponding to objects of statistics;Video acquisition is carried out to default statistical regions, monitor video is obtained;By Object identifying model, at least two field pictures included successively to monitor video are identified, and obtain the number of objects of statistics in each two field picture;Kth frame image is determined, and using the number of objects of statistics in kth frame image as the first quantity, wherein kth frame image is identified the image that the number including objects of statistics is not equal to zero for first;Judge whether to include in the i-th two field picture difference objects of statistics not to be covered in the i-th 1 two field pictures, if it is, the number of objects of statistics will be distinguished as corresponding second quantity of the i-th two field picture in the i-th two field picture, wherein i is the positive integer more than k;The flow of objects of statistics is determined according to the first quantity and each second quantity.This programme can reduce the cost for carrying out vehicle flowrate.
Description
Technical field
The present invention relates to information acquiring technology field, more particularly to a kind of flow statistical method and device.
Background technology
With the continuous improvement of living standard, private car has become the universal trip vehicles so that each department vapour
Car recoverable amount sustainable growth.With the continuous growth of car ownership, the pressure that road traffic is born also is continuously increased, caused
The trend increased is presented in the probability and the time of traffic congestion that traffic accident occurs.Vehicle supervision department is in order to understanding in real time
The jam situation of road traffic to vehicle flowrate, it is necessary to count.
Now, when being counted to vehicle flowrate, sensor generally is laid under pavement of road, such as in highway
Sensor is laid under road surface in gateway.When vehicle travels are equipped with the road surface of sensor, sensor can sense what is crossed
Automobile, so as to realize the statistics of flow motor.
The method that is counted for target to vehicle flowrate under pavement of road, it is necessary to lay sensor, and by sensor
The construction cost being routed under road surface is higher, causes the cost for carrying out vehicle flowrate higher.
The content of the invention
The embodiments of the invention provide a kind of flow statistical method and device, can reduce carry out vehicle flowrate into
This.
In a first aspect, the embodiments of the invention provide a kind of flow statistical method, object of the generation corresponding to objects of statistics
Identification model, in addition to:
Video acquisition is carried out to default statistical regions, monitor video is obtained;
By the Object identifying model, at least two field pictures that the monitor video includes are identified successively, obtained
Obtain the number of objects of statistics described in each frame described image;
From at least two field pictures determine kth frame image, and by objects of statistics described in the kth frame image
Count as the first quantity, wherein, the kth frame image is identified the number including the objects of statistics for first and is not equal to
Zero described image, and the k is positive integer;
For the i-th two field picture at least two field pictures, judge whether include i-th -1 in i-th two field picture
Difference objects of statistics not to be covered in two field picture, if it is, the number that objects of statistics is distinguished described in i-th two field picture is made
For corresponding second quantity of i-th two field picture, wherein, the i is the positive integer more than the k;
According to first quantity and each described second quantity, the flow of the objects of statistics is determined.
Alternatively,
The objects of statistics includes:Automobile, motorcycle, bicycle or pedestrian.
Alternatively,
The generation corresponds to the Object identifying model of objects of statistics, including:
Obtain at least one training image for including the objects of statistics;
On each Zhang Suoshu training images, objects of statistics described in the training image is carried out in the form of rectangle frame
Mark;
Convolutional neural networks model is trained using each Zhang Suoshu training images after being identified, acquisition can be known
The Object identifying model of not described objects of statistics.
Alternatively,
It is described by the Object identifying model, at least two field pictures that the monitor video includes are known successively
Not, the number of objects of statistics described in each frame described image is obtained, including:
Obtain at least two field pictures that the monitor video includes;
At least two field pictures are input in the Object identifying model successively, at least two field pictures
Each two field picture, rectangle frame is utilized respectively by the Object identifying model and identifies each described statistics in the two field picture
Object, the number for obtaining the rectangle frame exported by the Object identifying model is used as objects of statistics described in the two field picture
Number.
Alternatively,
It is described to judge whether to include in i-th two field picture difference objects of statistics not to be covered in the i-th -1 two field picture, bag
Include:
S1:Judge whether include the objects of statistics in i-th two field picture, if it is, performing S2, otherwise perform
S6;
S2:Judge whether include the objects of statistics in i-th -1 two field picture, if it is, performing S3, otherwise perform
S5;
S3:By the Object identifying model, i-th two field picture and i-th -1 frame figure are identified using rectangle frame
Each described objects of statistics as in, and obtain the coordinate information of each rectangle frame;
S4:According to the coordinate information of each rectangle frame, judging ought with the presence or absence of at least one on i-th two field picture
Preceding rectangle frame, wherein, the current rectangle frame and each rectangle frame on i-th -1 two field picture to intersect area equal
Less than area threshold set in advance, if it is, performing S5, S6 is otherwise performed;
S5:Determine to include difference objects of statistics not to be covered in i-th -1 two field picture in i-th two field picture, and
Terminate current process;
S6:Determine not include the difference objects of statistics in i-th two field picture.
Alternatively,
Described in each frame described image of acquisition after the number of objects of statistics, further comprise:
If the number of objects of statistics is zero described in each frame described image, the flow for determining the objects of statistics is
Zero.
Second aspect, the embodiment of the present invention additionally provides a kind of flow statistic device, including:Model generation unit, video
Collecting unit, object identification unit, object statistic unit and flow rate calculation unit;
The model generation unit, for generating the Object identifying model corresponding to objects of statistics;
The video acquisition unit, for carrying out video acquisition to default statistical regions, obtains monitor video;
The object identification unit, for the Object identifying model generated by the model generation unit, to institute
At least two field pictures that stating the monitor video of video acquisition unit acquisition includes are identified successively, obtain described in each frame
The number of objects of statistics described in image;
The object statistic unit, for being united described in each frame described image for being obtained according to the object identification unit
The number of object is counted, the determination kth frame image from least two field pictures, and will statistics pair described in the kth frame image
The number of elephant as the first quantity, wherein, the kth frame image is first and is identified and includes the number of the objects of statistics
It is not equal to zero described image, and the k is positive integer;And for for the i-th two field picture at least two field pictures, sentencing
Break and whether include difference objects of statistics not to be covered in the i-th -1 two field picture in i-th two field picture, if it is, by described i-th
The number of objects of statistics is distinguished described in two field picture as corresponding second quantity of i-th two field picture, wherein, the i is big
In the positive integer of the k;
The flow rate calculation unit, for first quantity got according to the object statistic unit and each institute
The second quantity is stated, the flow of the objects of statistics is determined.
Alternatively,
The model generation unit, for obtaining at least one training image for including the objects of statistics, each
On Zhang Suoshu training images, objects of statistics described in the training image is identified in the form of rectangle frame, and utilizes
Each Zhang Suoshu training images after line identifier are trained to convolutional neural networks model, and acquisition can recognize the objects of statistics
The Object identifying model.
Alternatively,
The object identification unit, for obtaining at least two field pictures that the monitor video includes, successively by described in extremely
Few two field pictures are input in the Object identifying model, for each two field picture at least two field pictures, by described
Object identifying model is utilized respectively rectangle frame and identifies each described objects of statistics in the two field picture, obtains by the object
Identification model output the rectangle frame number as objects of statistics described in the two field picture number.
Alternatively,
The object statistic unit includes:Update subelement;
The renewal subelement, for performing following steps:
S1:Judge whether include the objects of statistics in i-th two field picture, if it is, performing S2, otherwise perform
S6;
S2:Judge whether include the objects of statistics in i-th -1 two field picture, if it is, performing S3, otherwise perform
S5;
S3:By the Object identifying model, i-th two field picture and i-th -1 frame figure are identified using rectangle frame
Each described objects of statistics as in, and obtain the coordinate information of each rectangle frame;
S4:According to the coordinate information of each rectangle frame, judging ought with the presence or absence of at least one on i-th two field picture
Preceding rectangle frame, wherein, the current rectangle frame and each rectangle frame on i-th -1 two field picture to intersect area equal
Less than area threshold set in advance, if it is, performing S5, S6 is otherwise performed;
S5:Determine to include difference objects of statistics not to be covered in i-th -1 two field picture in i-th two field picture, and
Terminate current process;
S6:Determine not include the difference objects of statistics in i-th two field picture.
Statistical regions are carried out video acquisition and obtained after monitor video by flow statistical method and device that the present invention is provided,
By being identified with each two field picture that objects of statistics includes to monitor video successively to corresponding Object identifying model, obtain
The number of objects of statistics in each two field picture.First kth frame image for including objects of statistics is determined from each two field picture,
And obtain after the first quantity of objects of statistics in kth frame image, for each two field picture after kth frame image, if the frame
Image includes previous frame image difference objects of statistics not to be covered, regard the number for distinguishing objects of statistics as two field picture correspondence
The second quantity, calculate the second quantity corresponding to each two field picture after the first quantity and kth frame image and during with monitor video
Long business as objects of statistics flow.As can be seen here, the first quantity and each the second quantity and reflect in acquisition monitoring
Enter the quantity of the objects of statistics of statistical regions in video time, so as to the flow of counting statistics object accordingly, therefore
When automobile is as objects of statistics, it is only necessary to the video passed through by video capture device collection vehicle, passed without being laid under road surface
Sensor, reduces construction cost, so as to reduce the cost of vehicle flowrate.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are the present invention
Some embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
These accompanying drawings obtain other accompanying drawings.
Fig. 1 is a kind of flow chart for flow statistical method that one embodiment of the invention is provided;
Fig. 2 is a kind of flow chart for difference objects of statistics recognition methods that one embodiment of the invention is provided;
Fig. 3 is the flow chart for another flow statistical method that one embodiment of the invention is provided;
Fig. 4 is the schematic diagram of equipment where a kind of flow statistic device that one embodiment of the invention is provided;
Fig. 5 is a kind of schematic diagram for flow statistic device that one embodiment of the invention is provided;
Fig. 6 is the schematic diagram for another flow statistic device that one embodiment of the invention is provided.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
A part of embodiment of the present invention, rather than whole embodiments, based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained on the premise of creative work is not made, belongs to the scope of protection of the invention.
As shown in figure 1, the embodiments of the invention provide a kind of flow statistical method, this method may comprise steps of:
Step 101:Object identifying model of the generation corresponding to objects of statistics;
Step 102:Video acquisition is carried out to default statistical regions, monitor video is obtained;
Step 103:By the Object identifying model, at least two field pictures included successively to the monitor video are carried out
Identification, obtains the number of objects of statistics described in each frame described image;
Step 104:Kth frame image is determined from least two field pictures, and will be counted described in the kth frame image
The number of object as the first quantity, wherein, the kth frame image be first be identified including the objects of statistics
Number is not equal to zero described image, and the k is positive integer;
Step 105:For the i-th two field picture at least two field pictures, judge whether include in i-th two field picture
There is difference objects of statistics not to be covered in the i-th -1 two field picture, if it is, objects of statistics will be distinguished described in i-th two field picture
Number as corresponding second quantity of i-th two field picture, wherein, the i is positive integer more than the k;
Step 106:According to first quantity and each described second quantity, the flow of the objects of statistics is determined.
The embodiments of the invention provide a kind of flow statistical method, video acquisition is carried out to statistical regions and obtains monitor video
Afterwards, by being identified with each two field picture that objects of statistics includes to monitor video successively to corresponding Object identifying model,
Obtain the number of objects of statistics in each two field picture.First kth frame for including objects of statistics is determined from each two field picture
Image, and obtain after the first quantity of objects of statistics in kth frame image, for each two field picture after kth frame image, if
The two field picture includes previous frame image difference objects of statistics not to be covered, regard the number for distinguishing objects of statistics as the two field picture
Corresponding second quantity, calculate the second quantity corresponding to each two field picture after the first quantity and kth frame image and is regarded with monitoring
The business of frequency duration as objects of statistics flow.As can be seen here, the first quantity and each the second quantity and reflect in collection
Enter the quantity of the objects of statistics of statistical regions in the monitor video time, so as to the flow of counting statistics object accordingly, because
This is when automobile is as objects of statistics, it is only necessary to the video passed through by video capture device collection vehicle, without in road surface lower berth
If sensor, construction cost is reduced, so as to reduce the cost of vehicle flowrate.
Alternatively, objects of statistics can be automobile, motorcycle, bicycle or pedestrian.Such as, pair of the generation corresponding to automobile
After identification model, by the video passed through in expressway access or outlet collection automobile, can to expressway access or
The flow motor of outlet is counted;For another example, after generation is corresponding to the Object identifying model of pedestrian, by market, train
Stand, the video that the outlet in the place such as bus station or entrance collection pedestrian pass through, can be to places such as market, railway station, bus stations
Pedestrian's flow counted.
The difference of objects of statistics according to corresponding to Object identifying model, the flow statistical method that present aspect embodiment is provided can
For accounting automobile, motorcycle, bicycle and the flow of pedestrian, therefore the flow statistical method has stronger applicability.
Alternatively, as shown in figure 1,
When step 101 generation corresponds to the Object identifying model of objects of statistics, obtain at least one and include objects of statistics
Training image, objects of statistics is identified in the form of rectangle frame on each training image, after being identified
Each training image convolutional neural networks model is trained, acquisition the object of objects of statistics can be identified from image
Identification model.
Training image can be the photo for including objects of statistics shot by camera or be clapped by video camera
That takes the photograph includes the video of objects of statistics.When training image is photo, identified respectively on each photo by rectangle frame
Objects of statistics;When training image is video, objects of statistics is identified by rectangle frame on each two field picture of video respectively.Often
A training image can be included on one training image, multiple training images can also be included.Wherein, in order to ensure object know
The accuracy that objects of statistics is identified other model, when identifying objects of statistics by rectangle frame on training image, rectangle frame
For the minimum rectangle frame that the image of objects of statistics can be included.
In order to improve the accuracy that objects of statistics is identified Object identifying model, training image as far as possible with Statistical Area
The environmental factor in domain is identical.Such as, can be using automobile in road when needing to count the flow that highway is got on the car
The photo or video that are travelled on road trains Object identifying model.
Target detection model can be the convolutional neural networks model such as YOLO, Faster R-CNN, and these models are without people
Work sets up the feature of object, but utilizes convolutional neural networks, abstract, advanced features of the automatic study object in visual aspects.
For example, being trained using the training image for being identified with rectangle frame to Faster R-CNN models, according to accurate to statistical result
Property requirement, to Faster R-CNN models carry out correspondence number of times training, obtain can recognize that objects of statistics object know
Other model.By being trained acquisition Object identifying model to convolutional neural networks model so that Object identifying model can be with
The objects of statistics that faster rate identification goes out in image, it is ensured that the speed and accuracy rate that statistical model is identified.
Alternatively, as shown in figure 1,
When step 103 obtains the number of objects of statistics in each two field picture in monitor video by Object identifying model,
Each two field picture that monitor video includes is obtained, each two field picture that monitor video includes is input to Object identifying model successively
In.Each two field picture that monitor video includes is directed to, Object identifying model identifies every in the two field picture using rectangle frame
One objects of statistics, and count that the total quantity of rectangle frame is identified on the two field picture as objects of statistics in the two field picture
Number.
When generating Object identifying model, due to identifying the objects of statistics on training image, therefore generation by rectangle frame
Object identifying model adopt in a like fashion, each objects of statistics on each two field picture is identified using rectangle frame.For
In one rectangle frame of each objects of statistics correspondence on same two field picture, the two field picture, therefore the total quantity of rectangle frame is this
The number of objects of statistics on two field picture.Object identifying model is not only able to count the frame figure when a two field picture is identified
Can be specifically rectangle as the quantity of upper rectangle frame, additionally it is possible to obtain coordinate information of each rectangle frame in the two field picture
The frame upper left corner and the coordinate in the lower right corner, or can be the coordinate in the rectangle frame lower left corner and the upper right corner, believe for rectangle frame coordinate
The purposes of breath will will be described later.
Object identifying model identifies each objects of statistics in each two field picture by rectangle frame, when in a two field picture not
During including objects of statistics, then the number of rectangle frame is zero on the two field picture, correspondingly on the two field picture objects of statistics number
It is zero.Object identifying model counts the number of objects of statistics on each two field picture in this way, it is ensured that the statistics pair of acquisition
As the accuracy of number, and then ensure the accuracy of objects of statistics flow.
Alternatively, step 105 judges whether to include in the i-th two field picture difference not to be covered in the i-th -1 two field picture in Fig. 1
Objects of statistics, specifically, as shown in Fig. 2 step 105 can be realized by following sub-step:
Step 201:Judge whether include objects of statistics in the i-th two field picture, if it is, performing step 202, otherwise perform
Step 206.
As shown in figure 1, the number of objects of statistics in each two field picture is had confirmed in step 103, for kth frame figure
Each two field picture as after, if the number of objects of statistics is not equal to zero in the two field picture, step is performed for the two field picture
Rapid 202, if in the two field picture objects of statistics be equal to zero, for the two field picture perform step 206.
Step 202:Judge whether include objects of statistics in the i-th -1 two field picture, if it is, performing step 203, otherwise perform
Step 205.
As shown in figure 1, the number of objects of statistics in each two field picture is had confirmed in step 103, it is determined that the i-th frame
Include in image after objects of statistics, determine whether include objects of statistics in the i-th -1 two field picture, if in the i-th -1 two field picture
The number of objects of statistics is not equal to zero, it is necessary to determine whether that the i-th two field picture is differed with whether including in the i-th -1 two field picture
Objects of statistics, correspondingly perform step 203;If the number of objects of statistics is equal to zero in the i-th -1 two field picture, illustrate the i-th frame
All objects of statistics are all emerging in image, correspondingly perform step 205.
Step 203:By Object identifying model, each in the i-th image and the i-th -1 two field picture is identified using rectangle frame
Objects of statistics, and obtain the coordinate information of each rectangle frame.
As described above, after a two field picture is input into Object identifying model, Object identifying model utilizes rectangle collimation mark
Know and each objects of statistics in the two field picture, and the coordinate information of each rectangle frame returned.Get herein the i-th two field picture and
The coordinate information of each rectangle frame in i-th -1 two field picture, such as get each rectangle in the i-th two field picture and the i-th -1 two field picture
The upper left corner of frame and the coordinate in the lower right corner.
Step 204:According to the coordinate information of each rectangle frame, judge current with the presence or absence of at least one in the i-th two field picture
Rectangle frame, if it is, performing step 205, otherwise performs step 206.
Getting the seat of each rectangle frame on the coordinate information of each rectangle frame and the i-th -1 two field picture on the i-th two field picture
Mark after information, each rectangle frame on the i-th two field picture of traversal, judge the rectangle frame whether with each on the i-th -1 two field picture
The intersecting area of rectangle frame is both less than area threshold set in advance, if it is, regarding the rectangle frame as a current rectangle frame.
After each rectangle frame on the i-th two field picture is traveled through, if the quantity of current rectangle frame is not equal to zero, for the i-th two field picture
Step 205 is performed, otherwise step 206 is performed for the i-th two field picture.
Step 205:Determine to include in the i-th two field picture difference objects of statistics not to be covered in the i-th -1 two field picture, and terminate
Current process.
If the i-th two field picture includes objects of statistics, and does not include objects of statistics in the i-th -1 two field picture, then the i-th two field picture
In each objects of statistics be difference objects of statistics;If it is determined that there is at least one current rectangle frame on the i-th two field picture
Afterwards, illustrate to have identified the objects of statistics for not having occur or do not identify on the i-th -1 two field picture from the i-th two field picture, accordingly
Ground is, it is necessary to which the number of objects of statistics will be distinguished as the foundation of counting statistics object flow.
Step 206:Determine to include difference objects of statistics in the middle part of the i-th two field picture.
If it is determined that there is no objects of statistics in the i-th two field picture, illustrate to enter statistical regions without new objects of statistics.If
I-th two field picture is identical with included objects of statistics in the i-th -1 two field picture, or the objects of statistics that the i-th two field picture includes is
The objects of statistics that i-th -1 two field picture includes, then illustrate in the i-th two field picture each objects of statistics in each two field picture before
Counted, without repeating to count.
It may include because each two field picture that monitor video includes is continuously shot, therefore in adjacent two field pictures same
One objects of statistics, in order to avoid carrying out repeating statistics to objects of statistics, it is thus necessary to determine that the statistics pair that adjacent two field pictures include
As if no is identical objects of statistics.After being identified by rectangle frame to the objects of statistics on each two field picture, due to rectangle
Frame is the minimum rectangle frame that can be included objects of statistics, if being located at two rectangle frames in adjacent two field pictures respectively
Intersecting area be more than or equal to area threshold set in advance, illustrate the two rectangle frames correspond to same objects of statistics,
In the flow of counting statistics object, the objects of statistics is no longer calculated in latter two field picture.So, the phase of rectangle frame is passed through
Cross surface is accumulated to judge whether the objects of statistics in adjacent two field pictures in rear image was counted, for being counted
The objects of statistics crossed then no longer is carried out repeating statistics, it is ensured that the accuracy of traffic statistics is carried out to objects of statistics.
Alternatively, as shown in figure 1,
After step 103, if each two field picture that monitor video includes does not include objects of statistics, illustrate in collection prison
In the duration for controlling video, without objects of statistics by statistical regions, the flow for correspondingly determining objects of statistics is zero.
Exemplified by the vehicle flowrate on to highway is counted below, to traffic statistics side provided in an embodiment of the present invention
Method is described in further detail, as shown in figure 3, this method may comprise steps of:
Step 301:Generate the corresponding Object identifying model of automobile.
In an embodiment of the invention, the video that automobile is travelled on road, each frame figure that video is included are obtained
As a training image, automobile is identified on each training image by rectangle frame, each rectangle frame correspondence one
Automobile, rectangle frame is the minimum rectangle frame that corresponding automobile can be included.Utilize the training after each mark rectangle frame
Image is trained to Faster R-CNN models, and acquisition can identify the Object identifying model of automobile from image.
Step 302:Video acquisition is carried out to statistical regions, monitor video is obtained.
In an embodiment of the invention, camera is installed in the entrance of highway, by camera with specific angle
Spend and video acquisition is carried out to the entrance of highway, the angle of camera drives into the automobile of highway that can photograph as mark
It is accurate.The video acquisition of preset duration is carried out to the entrance of highway by camera, monitor video is obtained.Such as, at a high speed
Highway-entrance carries out the video acquisition of 24 hours, a length of 24h monitor video during acquisition.
Step 303:Initialize automobile number counter count=0.
Step 304:Each two field picture that monitor video includes is input to Object identifying model successively, until first includes
The kth frame image for having automobile is transfused to Object identifying model, and automobile quantity is counted according to the input results of Object identifying model
Device count is updated.
In an embodiment of the invention, obtain in monitor video each group of picture into set V, set V can be with table
It is shown as V={ xi| i=1,2,3 ..., N }, wherein, xiFor the i-th two field picture in monitor video, N is the totalframes of monitor video.
Such as, monitor video each second includes 25 two field pictures, then the monitor video of 24h length, which amounts to, includes 2,160,000 two field pictures, correspondingly N
=2160000.
Each two field picture is input in Object identifying model successively, Object identifying model is since the 1st two field picture, with rectangle
Automobile on frame identification image, what Object identifying model was exported if it there is automobile on image is not empty set, if on image
In the absence of automobile then Object identifying model output be empty set.The identification process of Object identifying model can be expressed as follows:
{(cj,bboxj)|bboxj=(b1,b2,b3,b4, j=1,2 ..., M)=F (xi)
Wherein, cjRepresent j-th of automobile on the i-th two field picture, bboxjRepresent the corresponding rectangle frame of the automobile, b1,b2Point
The rectangle frame top left co-ordinate, b is not represented3,b4The rectangle frame bottom right angular coordinate is represented, it is total that M represents that the i-th two field picture is got on the car
Number.
Since the 1st two field picture, if the { (c of Object identifying model F outputsj,bboxj) it is empty set, illustrate the two field picture
On there is not automobile, next two field picture is input in Object identifying model F by continuation, until kth frame image, by kth frame figure
After input object identification model F, the { (c of Object identifying model F outputsj,bboxj) be not empty set, then according to F (xk) it is defeated
Go out the number of rectangle frame in result, automobile number counter count is updated.
For example, since the 1st two field picture in 2,160,000 two field pictures, each two field picture is input into Object identifying model F successively
In, the output result that preceding 99 two field picture is input to after Object identifying model F is empty set, and the 100th two field picture is input to object knowledge
Output result after other model F is not empty set, according to F (x100) output result in rectangle frame number I perform count=
Count+I, is updated to automobile number counter count.
Step 305:Since the two field picture of kth+1, each two field picture is input to Object identifying model successively, obtained corresponding
Output result.
In an embodiment of the invention, it is first from determination kth frame image to include after the image of automobile, successively will
Each two field picture after kth frame image is input in Object identifying model, obtains corresponding output result.
For example, since the 101st two field picture, each two field picture by after is input in Object identifying model F successively, obtain
Corresponding output result.
Step 306:According to the corresponding output result of each two field picture, automobile number counter count is updated.
In an embodiment of the invention, for after kth frame image each two field picture (i.e. i=k+1, k+2 ...,
N), if the corresponding output result F (x of the i-th two field picturei) be empty set, then automobile number counter count is not updated;
If the corresponding output result F (x of the i-th two field picturei) be not empty set, then according to output result F (xi) and output result F (xi-1) in
The top left co-ordinate and bottom right angular coordinate of each rectangle frame, judge output result F (xi) it whether there is at least one current rectangle
Frame, wherein, current rectangle frame and output result F (xi-1) in any one rectangle frame intersecting area it is respectively less than set in advance
Area threshold, if current rectangle frame is present, is carried out more according to the quantity of current rectangle frame to automobile number counter count
Newly so that count=count+P, wherein P is the quantity of current rectangle frame in the i-th two field picture.
For example, after the 101st two field picture input object identification model F, there is 1 rectangle frame and the 100th frame figure in output result
(it is such as every frame figure as the intersecting area of each rectangle frame in corresponding I rectangle frame is respectively less than default area threshold
Image planes product 1/5), then perform count=count+1.So, from the 101st two field picture to the 2160000th two field picture, by each frame
The output result of image and the output result of previous frame image are compared, to be updated to automobile number counter count,
After the completion of performing aforesaid operations to the 2160000th two field picture, automobile number counter count is 55000, i.e., in acquisition monitoring video
24 hours in, there are 55,000 two automobiles to enter highway from the high speed crossing for being provided with camera.
Step 307:Flow motor is calculated according to automobile number counter count and preset duration.
In an embodiment of the invention, calculate the business of automobile number counter count and preset duration, as it is default when
The flow motor of the expressway access in long.
For example, calculating automobile number counter 55000 and preset duration 24h business, the vapour of the expressway access is obtained
Vehicle flowrate is about 2292/hour.
It should be noted that in each above-mentioned embodiment, coordinate system used in different two field pictures is the same coordinate system.
As shown in Figure 4, Figure 5, the embodiments of the invention provide a kind of flow statistic device.Device embodiment can be by soft
Part is realized, can also be realized by way of hardware or software and hardware combining.For hardware view, as shown in figure 4, being this hair
A kind of hardware structure diagram of equipment where the flow statistic device that bright embodiment is provided, except the processor shown in Fig. 4, internal memory,
Outside network interface and nonvolatile memory, the equipment in embodiment where device can also generally include other hardware,
Such as it is responsible for the forwarding chip of processing message.Exemplified by implemented in software, as shown in figure 5, being used as the dress on a logical meaning
Put, be to read corresponding computer program instructions in nonvolatile memory in internal memory by the CPU of equipment where it to transport
What row was formed.The flow statistic device that the present embodiment is provided, including:Model generation unit 501, video acquisition unit 502, object
Recognition unit 503, object statistic unit 504 and flow rate calculation unit 505;
Model generation unit 501, for generating the Object identifying model corresponding to objects of statistics;
Video acquisition unit 502, for carrying out video acquisition to default statistical regions, obtains monitor video;
Object identification unit 503, for the Object identifying model generated by model generation unit 501, to video acquisition
At least two field pictures that the monitor video that unit 502 is obtained includes are identified successively, obtain objects of statistics in each two field picture
Number;
Object statistic unit 504, for objects of statistics in each two field picture for being obtained according to object identification unit 503
Number, determines kth frame image from least two field pictures, and using the number of objects of statistics in kth frame image as the first quantity, its
In, kth frame image is first and is identified the image that the number including objects of statistics is not equal to zero, and k is positive integer;It is used in combination
Whether the i-th two field picture in at least two field pictures, judge to include in the i-th two field picture does not include in the i-th -1 two field picture
Difference objects of statistics, if it is, using in the i-th two field picture distinguish objects of statistics number be used as the i-th two field picture corresponding second
Quantity, wherein, i is the positive integer more than k;
Flow rate calculation unit 505, for the first quantity got according to object statistic unit 504 and each second number
Amount, determines the flow of objects of statistics.
Alternatively, as shown in figure 5,
Model generation unit 501, for obtaining at least one training image for including objects of statistics, in each Zhang Xunlian
On image, in the form of the rectangle frame to training image in objects of statistics be identified, and utilize each Zhang Xunlian after being identified
Image is trained to convolutional neural networks model, and acquisition can recognize the Object identifying model of objects of statistics.
Alternatively, as shown in figure 5,
Object identification unit 503, successively will at least two field pictures for obtaining at least two field pictures that monitor video includes
It is input in Object identifying model, for each two field picture at least two field pictures, square is utilized respectively by Object identifying model
Shape frame identifies each objects of statistics in the two field picture, obtains the number conduct of the rectangle frame exported by Object identifying model
The number of objects of statistics in the two field picture.
Alternatively, as shown in fig. 6,
Object statistic unit 504 includes:Update subelement 5041;
Subelement 5041 is updated, for performing following steps:
S1:Judge whether include objects of statistics in the i-th two field picture, if it is, performing S2, otherwise perform S6;
S2:Judge whether include objects of statistics in the i-th -1 two field picture, if it is, performing S3, otherwise perform S5;
S3:By Object identifying model, each system in the i-th two field picture and the i-th -1 two field picture is identified using rectangle frame
Object is counted, and obtains the coordinate information of each rectangle frame;
S4:According to the coordinate information of each rectangle frame, judge to whether there is at least one current rectangle on the i-th two field picture
Frame, wherein, current rectangle frame is respectively less than area set in advance with the area that intersects of each rectangle frame on the i-th -1 two field picture
Threshold value, if it is, performing S5, otherwise performs S6;
S5:Determine to include in the i-th two field picture difference objects of statistics not to be covered in the i-th -1 two field picture, and terminate current
Flow;
S6:Determine not including difference objects of statistics in the i-th two field picture.
The contents such as the information exchange between each unit, implementation procedure in said apparatus, due to implementing with the inventive method
Example is based on same design, and particular content can be found in the narration in the inventive method embodiment, and here is omitted.
The embodiment of the present invention additionally provides the execute instruction that is stored with a kind of computer-readable recording medium, the computer-readable recording medium, works as storage
Described in the computing device of controller during execute instruction, the storage control performs the flow system that each above-mentioned embodiment is provided
Meter method.
The embodiment of the present invention additionally provides a kind of storage control, including:Processor, memory and bus;
The memory is used to store execute instruction, and the processor is connected with the memory by the bus, when
During the storage control operation, the execute instruction of memory storage described in the computing device, so that the storage
Controller performs the flow statistical method that each above-mentioned embodiment is provided.
In summary, each embodiment of the invention is provided flow statistical method and device, at least with following beneficial effect
Really:
1st, in embodiments of the present invention, to statistical regions carry out video acquisition obtain monitor video after, by with statistics pair
As each two field picture included successively to monitor video to corresponding Object identifying model is identified, obtain in each two field picture
The number of objects of statistics.Determine that first includes the kth frame image of objects of statistics from each two field picture, and obtain kth frame
In image after the first quantity of objects of statistics, for each two field picture after kth frame image, if the two field picture includes
One two field picture difference objects of statistics not to be covered, using the number of difference objects of statistics as corresponding second quantity of the two field picture,
Calculate the second quantity corresponding to each two field picture after the first quantity and kth frame image and the business with monitor video duration as system
Count the flow of object.As can be seen here, the first quantity and each the second quantity and reflect to enter in acquisition monitoring video time
Enter the quantity of the objects of statistics of statistical regions, so as to the flow of counting statistics object accordingly, therefore statistics is used as in automobile
During object, it is only necessary to the video passed through by video capture device collection vehicle, without laying sensor under road surface, reduce and apply
Work cost, so as to reduce the cost of vehicle flowrate.
2nd, in embodiments of the present invention, according to corresponding to Object identifying model objects of statistics difference, present aspect embodiment
The flow statistical method of offer can be used for accounting automobile, motorcycle, bicycle and the flow of pedestrian, therefore the traffic statistics side
Method has stronger applicability.
3rd, in embodiments of the present invention, by being trained to the convolutional neural networks model such as YOLO, Faster R-CNN
Obtain Object identifying model so that the objects of statistics that Object identifying model can be gone out in image with faster rate identification, it is ensured that
The speed and accuracy rate that statistical model is identified.
4th, in embodiments of the present invention, Object identifying model can be identified the objects of statistics on image using rectangle frame
Come, so, judge objects of statistics in adjacent two field pictures in rear image whether by the intersecting area of rectangle frame
Counted, and then no longer carried out repeating statistics for the objects of statistics counted, it is ensured that flow objects of statistics
Measure statistical accuracy.
It should be noted that herein, such as first and second etc relational terms are used merely to an entity
Or operation makes a distinction with another entity or operation, and not necessarily require or imply exist between these entities or operation
Any this actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant be intended to it is non-
It is exclusive to include, so that process, method, article or equipment including a series of key elements not only include those key elements,
But also other key elements including being not expressly set out, or also include solid by this process, method, article or equipment
Some key elements.In the absence of more restrictions, the key element limited by sentence " including one ", is not arranged
Except also there is other identical factor in the process including the key element, method, article or equipment.
One of ordinary skill in the art will appreciate that:Realizing all or part of step of above method embodiment can pass through
Programmed instruction related hardware is completed, and foregoing program can be stored in the storage medium of embodied on computer readable, the program
Upon execution, the step of including above method embodiment is performed;And foregoing storage medium includes:ROM, RAM, magnetic disc or light
Disk etc. is various can be with the medium of store program codes.
It is last it should be noted that:Presently preferred embodiments of the present invention is the foregoing is only, the skill of the present invention is merely to illustrate
Art scheme, is not intended to limit the scope of the present invention.Any modification for being made within the spirit and principles of the invention,
Equivalent substitution, improvement etc., are all contained in protection scope of the present invention.
Claims (10)
1. a kind of flow statistical method, it is characterised in that generation corresponds to the Object identifying model of objects of statistics, in addition to:
Video acquisition is carried out to default statistical regions, monitor video is obtained;
By the Object identifying model, at least two field pictures that the monitor video includes are identified successively, obtain every
The number of objects of statistics described in one frame described image;
Kth frame image is determined from least two field pictures, and the number of objects of statistics described in the kth frame image is made
For the first quantity, wherein, the kth frame image is first and is identified the number including the objects of statistics and is not equal to zero
Described image, and the k is positive integer;
For the i-th two field picture at least two field pictures, judge whether include the i-th -1 frame figure in i-th two field picture
The difference objects of statistics not to be covered as in, if it is, regarding the number that objects of statistics is distinguished described in i-th two field picture as institute
Corresponding second quantity of the i-th two field picture is stated, wherein, the i is the positive integer more than the k;
According to first quantity and each described second quantity, the flow of the objects of statistics is determined.
2. according to the method described in claim 1, it is characterised in that
The objects of statistics includes:Automobile, motorcycle, bicycle or pedestrian.
3. according to the method described in claim 1, it is characterised in that Object identifying mould of the generation corresponding to objects of statistics
Type, including:
Obtain at least one training image for including the objects of statistics;
On each Zhang Suoshu training images, rower is entered to objects of statistics described in the training image in the form of rectangle frame
Know;
Convolutional neural networks model is trained using each Zhang Suoshu training images after being identified, acquisition can recognize institute
State the Object identifying model of objects of statistics.
4. method according to claim 3, it is characterised in that described by the Object identifying model, successively to described
At least two field pictures that monitor video includes are identified, and obtain the number of objects of statistics described in each frame described image, bag
Include:
Obtain at least two field pictures that the monitor video includes;
At least two field pictures are input in the Object identifying model successively, for every at least two field pictures
One two field picture, by the Object identifying model be utilized respectively rectangle frame identify in the two field picture each it is described statistics pair
As the number for obtaining the rectangle frame exported by the Object identifying model is used as of objects of statistics described in the two field picture
Number.
5. according to any described method in Claims 1-4, it is characterised in that it is described judge in i-th two field picture whether
Include difference objects of statistics not to be covered in the i-th -1 two field picture, including:
S1:Judge whether include the objects of statistics in i-th two field picture, if it is, performing S2, otherwise perform S6;
S2:Judge whether include the objects of statistics in i-th -1 two field picture, if it is, performing S3, otherwise perform S5;
S3:By the Object identifying model, identified using rectangle frame in i-th two field picture and i-th -1 two field picture
Each described objects of statistics, and obtain the coordinate information of each rectangle frame;
S4:According to the coordinate information of each rectangle frame, judge to whether there is at least one current square on i-th two field picture
Shape frame, wherein, the current rectangle frame is respectively less than with the area that intersects of each rectangle frame on i-th -1 two field picture
Area threshold set in advance, if it is, performing S5, otherwise performs S6;
S5:Determine to include difference objects of statistics not to be covered in i-th -1 two field picture in i-th two field picture, and terminate
Current process;
S6:Determine not include the difference objects of statistics in i-th two field picture.
6. according to any described method in Claims 1-4, it is characterised in that in each frame described image of acquisition
After the number of the objects of statistics, further comprise:
If the number of objects of statistics is zero described in each frame described image, the flow for determining the objects of statistics is zero.
7. a kind of flow statistic device, it is characterised in that including:Model generation unit, video acquisition unit, Object identifying list
Member, object statistic unit and flow rate calculation unit;
The model generation unit, for generating the Object identifying model corresponding to objects of statistics;
The video acquisition unit, for carrying out video acquisition to default statistical regions, obtains monitor video;
The object identification unit, for the Object identifying model generated by the model generation unit, is regarded to described
At least two field pictures that the monitor video that frequency collecting unit is obtained includes are identified successively, obtain each frame described image
Described in objects of statistics number;
The object statistic unit, for statistics pair described in each frame described image for being obtained according to the object identification unit
The number of elephant, determines kth frame image from least two field pictures, and by objects of statistics described in the kth frame image
Number as the first quantity, wherein, the kth frame image is first and is identified and includes the number of the objects of statistics
In zero described image, and the k is positive integer;And for for the i-th two field picture at least two field pictures, judging institute
State and whether include difference objects of statistics not to be covered in the i-th -1 two field picture in the i-th two field picture, if it is, by the i-th frame figure
The number of objects of statistics is distinguished as described in as corresponding second quantity of i-th two field picture, wherein, the i is more than institute
State k positive integer;
The flow rate calculation unit, for first quantity that is got according to the object statistic unit and each described the
Two quantity, determine the flow of the objects of statistics.
8. device according to claim 7, it is characterised in that
The model generation unit, for obtaining at least one training image for including the objects of statistics, in each institute
State on training image, objects of statistics described in the training image is identified in the form of rectangle frame, and utilize rower
Each Zhang Suoshu training images after knowledge are trained to convolutional neural networks model, and acquisition can recognize the institute of the objects of statistics
State Object identifying model.
9. device according to claim 8, it is characterised in that
The object identification unit, for obtaining at least two field pictures that the monitor video includes, successively at least two by described in
Two field picture is input in the Object identifying model, for each two field picture at least two field pictures, by the object
Identification model is utilized respectively rectangle frame and identifies each described objects of statistics in the two field picture, obtains by the Object identifying
Model output the rectangle frame number as objects of statistics described in the two field picture number.
10. according to any described device in claim 7 to 9, it is characterised in that the object statistic unit includes:Update
Subelement;
The renewal subelement, for performing following steps:
S1:Judge whether include the objects of statistics in i-th two field picture, if it is, performing S2, otherwise perform S6;
S2:Judge whether include the objects of statistics in i-th -1 two field picture, if it is, performing S3, otherwise perform S5;
S3:By the Object identifying model, identified using rectangle frame in i-th two field picture and i-th -1 two field picture
Each described objects of statistics, and obtain the coordinate information of each rectangle frame;
S4:According to the coordinate information of each rectangle frame, judge to whether there is at least one current square on i-th two field picture
Shape frame, wherein, the current rectangle frame is respectively less than with the area that intersects of each rectangle frame on i-th -1 two field picture
Area threshold set in advance, if it is, performing S5, otherwise performs S6;
S5:Determine to include difference objects of statistics not to be covered in i-th -1 two field picture in i-th two field picture, and terminate
Current process;
S6:Determine not include the difference objects of statistics in i-th two field picture.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710761698.2A CN107316462A (en) | 2017-08-30 | 2017-08-30 | A kind of flow statistical method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710761698.2A CN107316462A (en) | 2017-08-30 | 2017-08-30 | A kind of flow statistical method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107316462A true CN107316462A (en) | 2017-11-03 |
Family
ID=60177387
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710761698.2A Pending CN107316462A (en) | 2017-08-30 | 2017-08-30 | A kind of flow statistical method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107316462A (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108647587A (en) * | 2018-04-23 | 2018-10-12 | 腾讯科技(深圳)有限公司 | Demographic method, device, terminal and storage medium |
CN108986465A (en) * | 2018-07-27 | 2018-12-11 | 深圳大学 | A kind of method of vehicle Flow Detection, system and terminal device |
CN109285376A (en) * | 2018-08-09 | 2019-01-29 | 同济大学 | A kind of bus passenger flow statistical analysis system based on deep learning |
CN109615925A (en) * | 2019-01-03 | 2019-04-12 | 上海钧正网络科技有限公司 | Vehicle parking control method, device, system and server based on video monitoring |
CN109785632A (en) * | 2019-03-14 | 2019-05-21 | 济南浪潮高新科技投资发展有限公司 | A kind of magnitude of traffic flow statistical method and device |
CN109859468A (en) * | 2019-01-30 | 2019-06-07 | 淮阴工学院 | Multilane traffic volume based on YOLOv3 counts and wireless vehicle tracking |
CN109902559A (en) * | 2019-01-15 | 2019-06-18 | 深圳中兴网信科技有限公司 | Statistical method of traffic flow, system and computer readable storage medium |
CN110021174A (en) * | 2019-04-02 | 2019-07-16 | 北京同方软件有限公司 | A kind of vehicle flowrate calculation method for being applicable in more scenes based on video image |
CN110287907A (en) * | 2019-06-28 | 2019-09-27 | 北京海益同展信息科技有限公司 | A kind of method for checking object and device |
CN110390822A (en) * | 2019-05-31 | 2019-10-29 | 东南大学 | Bridge statistical method of traffic flow based on FBG sensor and convolutional neural networks |
CN110738137A (en) * | 2019-09-26 | 2020-01-31 | 中移物联网有限公司 | people flow rate statistical method and device |
CN110853353A (en) * | 2019-11-18 | 2020-02-28 | 山东大学 | Vision-based density traffic vehicle counting and traffic flow calculating method and system |
CN111383455A (en) * | 2020-03-11 | 2020-07-07 | 上海眼控科技股份有限公司 | Traffic intersection object flow statistical method, device, computer equipment and medium |
CN112115792A (en) * | 2020-08-14 | 2020-12-22 | 深圳大学 | Access & exit information statistical system, method and computer equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103310444A (en) * | 2013-05-21 | 2013-09-18 | 上海交通大学 | Method of monitoring pedestrians and counting based on overhead camera |
CN104008396A (en) * | 2014-05-22 | 2014-08-27 | 南京邮电大学 | In and out people flow statistical method based on people head color and shape features |
CN104036639A (en) * | 2014-06-20 | 2014-09-10 | 上海理工大学 | Traffic flow statistics method |
CN106372570A (en) * | 2016-08-19 | 2017-02-01 | 云赛智联股份有限公司 | Visitor flowrate statistic method |
CN106874827A (en) * | 2015-12-14 | 2017-06-20 | 北京奇虎科技有限公司 | Video frequency identifying method and device |
-
2017
- 2017-08-30 CN CN201710761698.2A patent/CN107316462A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103310444A (en) * | 2013-05-21 | 2013-09-18 | 上海交通大学 | Method of monitoring pedestrians and counting based on overhead camera |
CN104008396A (en) * | 2014-05-22 | 2014-08-27 | 南京邮电大学 | In and out people flow statistical method based on people head color and shape features |
CN104036639A (en) * | 2014-06-20 | 2014-09-10 | 上海理工大学 | Traffic flow statistics method |
CN106874827A (en) * | 2015-12-14 | 2017-06-20 | 北京奇虎科技有限公司 | Video frequency identifying method and device |
CN106372570A (en) * | 2016-08-19 | 2017-02-01 | 云赛智联股份有限公司 | Visitor flowrate statistic method |
Non-Patent Citations (1)
Title |
---|
彭智: "智能交通系统中基于机器视觉的交通流量统计研究", 《中国优秀硕士学位论文全文数据库(电子期刊)》 * |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108647587A (en) * | 2018-04-23 | 2018-10-12 | 腾讯科技(深圳)有限公司 | Demographic method, device, terminal and storage medium |
CN108647587B (en) * | 2018-04-23 | 2021-08-24 | 腾讯科技(深圳)有限公司 | People counting method, device, terminal and storage medium |
CN108986465A (en) * | 2018-07-27 | 2018-12-11 | 深圳大学 | A kind of method of vehicle Flow Detection, system and terminal device |
CN109285376A (en) * | 2018-08-09 | 2019-01-29 | 同济大学 | A kind of bus passenger flow statistical analysis system based on deep learning |
CN109285376B (en) * | 2018-08-09 | 2022-04-19 | 同济大学 | Bus passenger flow statistical analysis system based on deep learning |
CN109615925A (en) * | 2019-01-03 | 2019-04-12 | 上海钧正网络科技有限公司 | Vehicle parking control method, device, system and server based on video monitoring |
CN109902559A (en) * | 2019-01-15 | 2019-06-18 | 深圳中兴网信科技有限公司 | Statistical method of traffic flow, system and computer readable storage medium |
CN109859468A (en) * | 2019-01-30 | 2019-06-07 | 淮阴工学院 | Multilane traffic volume based on YOLOv3 counts and wireless vehicle tracking |
CN109785632A (en) * | 2019-03-14 | 2019-05-21 | 济南浪潮高新科技投资发展有限公司 | A kind of magnitude of traffic flow statistical method and device |
CN109785632B (en) * | 2019-03-14 | 2021-05-04 | 浪潮集团有限公司 | Traffic flow statistical method and device |
CN110021174A (en) * | 2019-04-02 | 2019-07-16 | 北京同方软件有限公司 | A kind of vehicle flowrate calculation method for being applicable in more scenes based on video image |
CN110390822A (en) * | 2019-05-31 | 2019-10-29 | 东南大学 | Bridge statistical method of traffic flow based on FBG sensor and convolutional neural networks |
CN110287907A (en) * | 2019-06-28 | 2019-09-27 | 北京海益同展信息科技有限公司 | A kind of method for checking object and device |
WO2020258978A1 (en) * | 2019-06-28 | 2020-12-30 | 北京海益同展信息科技有限公司 | Object detection method and device |
CN110287907B (en) * | 2019-06-28 | 2020-11-03 | 北京海益同展信息科技有限公司 | Object detection method and device |
CN110738137A (en) * | 2019-09-26 | 2020-01-31 | 中移物联网有限公司 | people flow rate statistical method and device |
CN110853353B (en) * | 2019-11-18 | 2020-10-16 | 山东大学 | Vision-based density traffic vehicle counting and traffic flow calculating method and system |
CN110853353A (en) * | 2019-11-18 | 2020-02-28 | 山东大学 | Vision-based density traffic vehicle counting and traffic flow calculating method and system |
CN111383455A (en) * | 2020-03-11 | 2020-07-07 | 上海眼控科技股份有限公司 | Traffic intersection object flow statistical method, device, computer equipment and medium |
CN112115792A (en) * | 2020-08-14 | 2020-12-22 | 深圳大学 | Access & exit information statistical system, method and computer equipment |
CN112115792B (en) * | 2020-08-14 | 2024-02-06 | 深圳大学 | Entrance and exit information statistics system, method and computer equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107316462A (en) | A kind of flow statistical method and device | |
CN110364008B (en) | Road condition determining method and device, computer equipment and storage medium | |
TWI448977B (en) | Method and apparatus for video analytics based object counting | |
CN104537841B (en) | Unlicensed vehicle violation detection method and detection system thereof | |
CN109584553A (en) | A kind of section degree of association missing complementing method based on space time information | |
CN107301776A (en) | Track road conditions processing and dissemination method based on video detection technology | |
WO2022037387A1 (en) | Visual perception algorithm evaluation method and device | |
CN109064495A (en) | A kind of bridge floor vehicle space time information acquisition methods based on Faster R-CNN and video technique | |
CN106571038A (en) | Method for fully automatically monitoring road | |
CN110298300A (en) | A method of detection vehicle violation crimping | |
CN104882005B (en) | A kind of method and device for detecting lane flow amount | |
CN101882374A (en) | System and method for analysis of comprehensive information on vehicle trip distribution | |
CN110032947A (en) | A kind of method and device that monitor event occurs | |
CN106971544A (en) | A kind of direct method that vehicle congestion is detected using still image | |
WO2021036243A1 (en) | Method and apparatus for recognizing lane, and computing device | |
CN107316332A (en) | The camera and scene relating scaling method and system of a kind of application intelligent driving | |
CN106548629A (en) | Traffic violation detection method and system based on data fusion | |
CN107316463A (en) | A kind of method and apparatus of vehicle monitoring | |
CN110147731A (en) | Vehicle type recognition method and Related product | |
CN117037084A (en) | Method and device for identifying abnormal behavior of toll station vehicle based on track reconstruction | |
CN117456482B (en) | Abnormal event identification method and system for traffic monitoring scene | |
CN112488042B (en) | Pedestrian traffic bottleneck discrimination method and system based on video analysis | |
Suttiponpisarn et al. | Detection of wrong direction vehicles on two-way traffic | |
CN103473925A (en) | Verification method of road vehicle detection system | |
CN109720353A (en) | A kind of driving behavior detection method based on smart phone |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171103 |
|
RJ01 | Rejection of invention patent application after publication |