CN107689158A - A kind of intellectual traffic control method based on image procossing - Google Patents
A kind of intellectual traffic control method based on image procossing Download PDFInfo
- Publication number
- CN107689158A CN107689158A CN201710683208.1A CN201710683208A CN107689158A CN 107689158 A CN107689158 A CN 107689158A CN 201710683208 A CN201710683208 A CN 201710683208A CN 107689158 A CN107689158 A CN 107689158A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- pedestrian
- control method
- method based
- traffic control
- 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
- G08G1/07—Controlling traffic signals
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/07—Controlling traffic signals
- G08G1/08—Controlling traffic signals according to detected number or speed of vehicles
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a kind of intellectual traffic control method based on image procossing, comprise the following steps：Obtain the vedio data at current crossing；Analyze and process vedio data；Vedio data is differentiated；Passing rules are generated, issue current order.The present invention is detected and transmitted to traffic related information in a manner of image, and passing rules are generated according to actual traffic status, traffic can intuitively be judged, the current order of issue exactly, control traffic intelligent and high-efficiency, the traffic efficiency of pedestrian and vehicle is improved, the possibility to cause obstruction to traffic is reduced, meets the current requirement of pedestrian and vehicle.
Description
Technical field
The present invention relates to intelligent transportation field, especially a kind of intellectual traffic control method based on image procossing.
Background technology
Now, with the continuous social and economic development, the quality of life of people there has also been great raising, and many families are several
There is the private car of oneself, the number of trip also gradually increases, and this is one for the traffic lights on road and huge examined
Test.
At present, almost the light on and off time of most of traffic lights and translative mode are all fixed, and this causes people to pass through
The situation of traffic jam can be often run into, which part reason is that traffic lights can not provide current side well under some road conditions
Case, such as, performing the scheme that provides of traffic lights in some cases can to occur a direction traffic jam and other direction
There is no the situation of vehicle, specifically, the big road of vehicle flowrate just green time is short, while when pedestrian measures big road red light
Between it is short, traffic lights seems excessively inflexible in this case, and traffic lights can not be according to vehicle flowrate actual on road and pedestrian's number pair
Passing rules make appropriate change, it is impossible to effective intelligent control traffic in real time, therefore traffic smooth can not efficiently be run,
Traffic jam is easily caused, current trouble is brought to vehicle and pedestrian.
The content of the invention
In order to solve the above problems, it is an object of the invention to provide a kind of intellectual traffic control side based on image procossing
Method, according to road traffic condition real-time release passing rules, so as to intelligent control traffic lights.
To solve the above problems, the technical solution adopted by the present invention is：
A kind of intellectual traffic control method based on image procossing, comprises the following steps：
A, the vedio data at current crossing is obtained；
B, the vedio data is analyzed and processed；
C, the vedio data is differentiated；
D, passing rules are generated, issue current order.
Further, the vedio data at current crossing is obtained in step A, is comprised the following steps：
A1, the traffic video stream at the current crossing of collection；
A2, the automobile video frequency stream in traffic video stream and pedestrian's video flowing separated.
Further, vedio data is analyzed and processed in step B to regard including the vehicle in detection automobile video frequency stream and pedestrian
Pedestrian in frequency stream.
Further, detecting the vehicle in automobile video frequency stream includes the vehicle in automobile video frequency stream is counted and identified,
And generate information of vehicle flowrate and class of vehicle information.
Further, detecting the pedestrian in pedestrian's video flowing includes counting the pedestrian in pedestrian's video flowing, and generates
Walk amount information.
Further, vedio data is differentiated in step C, comprised the following steps：
C1, predetermined passing rules are generated according to information of vehicle flowrate, class of vehicle information and walk amount information；
C2, with reference to current crossing prevailing state information and predetermined passing rules, generate current decision-making.
Further, the vehicle in automobile video frequency stream count and realized by deep learning method, specific steps
Including：
B11, the fast convolution neutral net based on region is created, the fast convolution neutral net includes convolutional layer, pond
Change layer, full articulamentum and screening layer；
B12, select candidate region from the target image at current crossing；
B13, using convolutional layer and pond layer target image is handled, produce the characteristic pattern of candidate region；
B14, pond layer extract characteristic vector from the characteristic pattern of candidate region；
B15, characteristic vector is sent into full articulamentum, full articulamentum exports two output layers at the same level, one of output
The probabilistic estimated value of layer output vehicle, the bounding box of another output layer output token vehicle；
B16, bounding box is sent into screening layer, vehicle is counted.
Further, the pedestrian in pedestrian's video flowing count and realized by deep learning method, specific steps
Including：
B21, visual angle figure is obtained from target image；
B22, estimation visual angle figure and on the basis of the figure of visual angle according to the pedestrian position at candidate region center create crowd it is close
Degree figure；
B23, matching network is established, the matching network is trained using crowd density figure as true value；
B24, the localized mass with recognizable visual angle, yardstick and Density Distribution is retrieved from target image；
B25, target image is input in matching network, obtains pedestrian density's distribution map；
B26, point similar to the visual angle of the localized mass, yardstick and Density Distribution is retrieved from pedestrian density's distribution map
Block；
B27, according to piecemeal generate convolutional neural networks, pedestrian is counted by the convolutional neural networks.
Further, the vehicle in automobile video frequency stream is identified including identification ambulance, police car and fire fighting truck.
Further, the specific steps vehicle in automobile video frequency stream being identified include：
B11, establish a database relevant with police car, fire fighting truck and ambulance；
B12, a part for database is used as to training set, another part is used as test set；
B13, using training set the vehicle in the video that captures is identified.
The beneficial effects of the invention are as follows：A kind of intellectual traffic control method based on image procossing provided by the invention, it is first
The vedio data at current crossing is first obtained, then analyzes and processes vedio data, it is then pair relevant with vedio data
Information differentiated, ultimately produce passing rules, issue current order, the present invention is in a manner of image to traffic related information
Detected and transmitted, and passing rules are generated according to actual traffic status, can intuitively judged traffic, send out exactly
The current order of cloth, controls traffic intelligent and high-efficiency, improves the traffic efficiency of pedestrian and vehicle, reduce the possibility to cause obstruction to traffic
Property, meet the current requirement of pedestrian and vehicle.
Brief description of the drawings
Present pre-ferred embodiments are provided below in conjunction with the accompanying drawings, to describe embodiment of the present invention in detail.
Fig. 1 is the key step figure of the intellectual traffic control method of the present invention；
Fig. 2 is the flow chart of a preferred embodiment of the intellectual traffic control method of the present invention.
Embodiment
A kind of reference picture 1, intellectual traffic control method based on image procossing, comprises the following steps：
A, the vedio data at current crossing is obtained；
B, the vedio data is analyzed and processed；
C, the vedio data is differentiated；
D, passing rules are generated, issue current order.
Specifically, traffic related information is detected and transmitted in a manner of image, and given birth to according to actual traffic status
Into passing rules, traffic can be intuitively judged, the current order of issue, controls traffic intelligent and high-efficiency exactly, improves
Pedestrian and the traffic efficiency of vehicle, reduce the possibility to cause obstruction to traffic, meet the current requirement of pedestrian and vehicle.
Wherein, reference picture 2, the vedio data at current crossing is obtained in step A, is comprised the following steps：
A1, the traffic video stream at the current crossing of collection；
A2, the automobile video frequency stream in traffic video stream and pedestrian's video flowing separated.
It may contain other specific video datas in traffic video stream, such as number on road, the shop in road roadside etc.,
Therefore need to be filtered out, to obtain more preferable Detection results.
Wherein, referring to Figures 1 and 2, vedio data is analyzed and processed in step B to be included detecting the car in automobile video frequency stream
And pedestrian's video flowing in pedestrian.
Wherein, referring to Figures 1 and 2, detecting the vehicle in automobile video frequency stream includes carrying out the vehicle in automobile video frequency stream
Count and identify, and generate information of vehicle flowrate and class of vehicle information.
Wherein, referring to Figures 1 and 2, detecting the pedestrian in pedestrian's video flowing includes carrying out the pedestrian in pedestrian's video flowing
Count, and generate walk amount information.
Wherein, reference picture 1, vedio data is differentiated in step C, comprised the following steps：
C1, predetermined passing rules are generated according to information of vehicle flowrate, class of vehicle information and walk amount information；
C2, with reference to current crossing prevailing state information and predetermined passing rules, generate current order.
Specifically, predetermined passing rules are not necessarily the rule finally issued, that is to say, that with reference to the current shape in current crossing
State information and predetermined passing rules consider, can carry out comprehensive descision to produce the current decision-making of an adaptation, it is probably predetermined
The improvement of passing rules, it is also possible to the predetermined passing rules generated at the very start.
Wherein, the vehicle in automobile video frequency stream count and realized by deep learning method, specific steps bag
Include：
B11, the fast convolution neutral net based on region is created, the fast convolution neutral net includes convolutional layer, pond
Change layer, full articulamentum and screening layer；
B12, select candidate region from the target image at current crossing；
B13, using convolutional layer and pond layer target image is handled, produce the characteristic pattern of candidate region；
B14, pond layer extract characteristic vector from the characteristic pattern of candidate region；
B15, characteristic vector is sent into full articulamentum, full articulamentum exports two output layers at the same level, one of output
The probabilistic estimated value of layer output vehicle, the bounding box of another output layer output token vehicle；
B16, bounding box is sent into screening layer, vehicle is counted.
Specifically, the fast convolution neutral net based on region is an object detection network, its can by target image and
One or more candidate regions in target image as input, using caused bounding box and corresponding object estimate classification as
Output, can utilize bottom-up algorithms selection candidate region, bottom-up algorithm is a kind of publicly-owned technology, bottom-up
Algorithm be according to system functional requirement, since specific device, logical block or similar system, by carrying out phase to it
Connect, change and expand, form a kind of algorithm of required system, be used as bottom base by the use of target image in the method
Plinth, the candidate field of selection is upper strata basis, it is also possible to which selective search method produces candidate region, selective search method
It is to be grouped based on layering to establish a kind of data algorithm of model, the optimization more directly perceived compared with original exhaust algorithm, wherein
Selective search method it is conventional have two kinds, both approaches are prior art, are accelerated model and characteristic model respectively, separately
On the one hand, stochastic gradient descent method fine setting fast convolution neutral net can be utilized typically after full articulamentum is divided into two layers of output,
Stochastic gradient descent method, it is a kind of existing algorithm, abbreviation SPGD algorithms, as a kind of model-free optimized algorithm, is often applied to
Control variable is more, and controlled system is more complicated, can not establish the optimized control process of accurate mathematical model, last screening layer
With the screening rule defined in advance, i.e., the rule formulated according to the probability Estimation layer of output is specific next to screen bounding box
Say, i.e., screen bounding box using the Duplication of probability, size, length-width ratio and other bounding boxes, so as to vehicle count.
Wherein, the pedestrian in pedestrian's video flowing count and realized by deep learning method, specific steps bag
Include：
B21, visual angle figure is obtained from target image；
B22, estimation visual angle figure and on the basis of the figure of visual angle according to the pedestrian position at candidate region center create crowd it is close
Degree figure；
B23, matching network is established, the matching network is trained using crowd density figure as true value；
B24, the localized mass with recognizable visual angle, yardstick and Density Distribution is retrieved from target image；
B25, target image is input in matching network, obtains pedestrian density's distribution map；
B26, point similar to the visual angle of the localized mass, yardstick and Density Distribution is retrieved from pedestrian density's distribution map
Block；
B27, according to piecemeal generate convolutional neural networks, pedestrian is counted by the convolutional neural networks.
Specifically, the corresponding special scenes of the piecemeal similar with localized mass visual angle, yardstick and Density Distribution, further generation is special
Determine the convolutional neural networks under scene, the statistical information of pedestrian's amount can be obtained from this network.
Wherein, reference picture 2, the vehicle in automobile video frequency stream is identified including identification ambulance, police car and fire fighting truck,
Meet the principle of ambulance, police car and fire fighting truck priority pass, moreover, the vehicle of identification also include other public offices office or
Optionally it is badly in need of current vehicle.
Wherein, the specific steps vehicle in automobile video frequency stream being identified include：
B11, establish a database relevant with police car, fire fighting truck and ambulance；
B12, a part for database is used as to training set, another part is used as test set；
B13, using training set the vehicle in the video that captures is identified.
Specifically, test set coordinates training set to realize retrieval to database together, so as to filter out police car therein, fire-fighting
The classification information of car and ambulance, to be further identified by training set.
It is below a preferred embodiment of intellectual traffic control method of the invention.
For a thing and the intersection in north and south, specifically, if the transport information obtained after analyzing and processing is thing
Direction average vehicle flow is more than North and South direction average vehicle flow, and the average pedestrian amount of east-west direction is less than the average row of North and South direction
People measures, and the quantity of the police car of east-west direction, fire fighting truck and ambulance is more than the number of the police car of North and South direction, fire fighting truck and ambulance
Amount, the then predetermined passing rules generated may be：The current order of east-west direction wagon flow is current sequentially prior to North and South direction wagon flow,
In particular cases, when the vehicle number that east-west direction detects is identical with vehicle number that North and South direction detects, or thing and
When North and South direction has police car, fire fighting truck or ambulance, the vehicle of east-west direction is still prior to the vehicle pass-through of North and South direction.
Or using an actual numerical value as boundary, the predetermined passing rules of generation may be：If east-west direction pavement
Pedestrian's number be more than 15 or North and South direction driveway vehicle number be more than ten when, then allow east-west direction pavement or thing
Direction driveway passes through, and forbids North and South direction vehicle pass-through or east-west direction vehicle pass-through, it should be noted that this section of reality
Apply the numerical value " 15 " in example or " ten " do not characterize the design parameter of the present invention, be intended merely to be better described the work of the present invention
Make principle.
Or the quantity of Main Basiss fire fighting truck, police car and ambulance is handled, the predetermined passing rules of generation may
For：If only having in thing and North and South direction on a section when having fire fighting truck, police car or ambulance, fire fighting truck, police car or rescue
Section priority pass where protecting car；If have fire fighting truck, police car or ambulance in thing and North and South direction, it can set common
Current order between vehicle and above-mentioned three kinds of special cars, such as police car, fire fighting truck, the priority pass of ambulance and common vehicle
Sequentially it is：Fire fighting truck>Ambulance>Police car>Common vehicle.
In addition, reference picture 2, in another embodiment of the present invention, the passing rules that are just provided originally on traffic lights
Traditional rule is can be considered, traditional rule can form complement mode with predetermined passing rules, occur in predetermined passing rules larger inclined
When difference or some unpredictable failures of generation, then predetermined passing rules are not performed, mainly perform traditional rule, to ensure intelligence
The operation of energy traffic lights safety, such as, some unpredictable failures occur causes the traffic lights in two directions that intersect same
Shi Liangqi, predetermined passing rules typically at this moment not being considered mainly, but mainly performing the rule on original traffic lights, one kind is current
Rule is, the traffic lights of four direction passes through one minute successively according to thing and North and South direction, and this current mode at least can be with
Ensure that road is sane safely.
Presently preferred embodiments of the present invention and general principle are discussed in detail above content, but the invention is not limited in
Above-mentioned embodiment, those skilled in the art should be recognized that also had on the premise of without prejudice to spirit of the invention it is various
Equivalent variations and replacement, these equivalent variations and replacement all fall within the protetion scope of the claimed invention.
Claims (10)
- A kind of 1. intellectual traffic control method based on image procossing, it is characterised in that comprise the following steps：A, the vedio data at current crossing is obtained；B, the vedio data is analyzed and processed；C, the vedio data is differentiated；D, passing rules are generated, issue current order.
- A kind of 2. intellectual traffic control method based on image procossing according to claim 1, it is characterised in that the step The vedio data at current crossing is obtained in rapid A, is comprised the following steps：A1, the traffic video stream at the current crossing of collection；A2, the automobile video frequency stream in the traffic video stream and pedestrian's video flowing separated.
- A kind of 3. intellectual traffic control method based on image procossing according to claim 1, it is characterised in that the step The vedio data is analyzed and processed in rapid B to be included detecting in the vehicle and pedestrian's video flowing in the automobile video frequency stream Pedestrian.
- A kind of 4. intellectual traffic control method based on image procossing according to claim 3, it is characterised in that detection institute Stating the vehicle in automobile video frequency stream includes the vehicle in the automobile video frequency stream is counted and identified, and generates vehicle flowrate letter Breath and class of vehicle information.
- A kind of 5. intellectual traffic control method based on image procossing according to claim 4, it is characterised in that detection institute Stating the pedestrian in pedestrian's video flowing includes counting the pedestrian in pedestrian's video flowing, and generates walk amount letter Breath.
- A kind of 6. intellectual traffic control method based on image procossing according to claim 5, it is characterised in that the step The vedio data is differentiated in rapid C, comprised the following steps：C1, make a reservation for pass through according to the generation of the information of vehicle flowrate, the class of vehicle information and the walk amount information and advise Then；C2, with reference to current crossing prevailing state information and the predetermined passing rules, generate current decision-making.
- 7. a kind of intellectual traffic control method based on image procossing according to claim 5, it is characterised in that to described Vehicle in automobile video frequency stream count and realized by deep learning method, and specific steps include：B11, create the fast convolution neutral net based on region, the fast convolution neutral net include convolutional layer, pond layer, Full articulamentum and screening layer；B12, select candidate region from the target image at current crossing；B13, using convolutional layer and pond layer target image is handled, produce the characteristic pattern of candidate region；B14, pond layer extract characteristic vector from the characteristic pattern of candidate region；B15, characteristic vector is sent into full articulamentum, full articulamentum exports two output layers at the same level, and one of output layer is defeated Go out the probabilistic estimated value of vehicle, the bounding box of another output layer output token vehicle；B16, bounding box is sent into screening layer, vehicle is counted.
- 8. a kind of intellectual traffic control method based on image procossing according to claim 7, it is characterised in that to described Pedestrian in pedestrian's video flowing count and realized by deep learning method, and specific steps include：B21, visual angle figure is obtained from target image；B22, estimation visual angle figure simultaneously create crowd density figure on the basis of the figure of visual angle according to the pedestrian position at candidate region center；B23, matching network is established, the matching network is trained using crowd density figure as true value；B24, the localized mass with recognizable visual angle, yardstick and Density Distribution is retrieved from target image；B25, target image is input in matching network, obtains pedestrian density's distribution map；B26, the piecemeal similar to the visual angle of the localized mass, yardstick and Density Distribution is retrieved from pedestrian density's distribution map；B27, according to piecemeal generate convolutional neural networks, pedestrian is counted by the convolutional neural networks.
- 9. a kind of intellectual traffic control method based on image procossing according to claim 4, it is characterised in that to described Vehicle in automobile video frequency stream is identified including identification ambulance, police car and fire fighting truck.
- 10. a kind of intellectual traffic control method based on image procossing according to claim 9, it is characterised in that to institute Stating the specific steps that the vehicle in automobile video frequency stream is identified includes：B11, establish a database relevant with police car, fire fighting truck and ambulance；B12, a part for database is used as to training set, another part is used as test set；B13, using training set the vehicle in the video that captures is identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710683208.1A CN107689158A (en) | 2017-08-10 | 2017-08-10 | A kind of intellectual traffic control method based on image procossing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710683208.1A CN107689158A (en) | 2017-08-10 | 2017-08-10 | A kind of intellectual traffic control method based on image procossing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107689158A true CN107689158A (en) | 2018-02-13 |
Family
ID=61153258
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710683208.1A Pending CN107689158A (en) | 2017-08-10 | 2017-08-10 | A kind of intellectual traffic control method based on image procossing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107689158A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108961782A (en) * | 2018-08-21 | 2018-12-07 | 北京深瞐科技有限公司 | Traffic intersection control method and device |
| CN110047272A (en) * | 2019-05-21 | 2019-07-23 | 重庆工程学院 | A kind of intelligent transportation pedestrian behavior monitoring and alarming system based on big data |
| CN110363989A (en) * | 2019-07-11 | 2019-10-22 | 汉王科技股份有限公司 | Magnitude of traffic flow detection method, device, electronic equipment and storage medium |
| CN110826456A (en) * | 2019-10-31 | 2020-02-21 | 青岛海信网络科技股份有限公司 | Countdown board fault detection method and system |
| CN111746728A (en) * | 2020-06-17 | 2020-10-09 | 重庆大学 | Novel overwater cleaning robot based on reinforcement learning and control method |
| CN112071070A (en) * | 2020-09-18 | 2020-12-11 | 广州维他科技有限公司 | Intelligent traffic intersection controller |
| CN112991782A (en) * | 2021-04-08 | 2021-06-18 | 河北工业大学 | Control method, system, terminal, equipment, medium and application of traffic signal lamp |
| CN113781784A (en) * | 2021-11-09 | 2021-12-10 | 深圳市奥新科技有限公司 | Intelligent traffic light and control method thereof |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101577054A (en) * | 2009-02-27 | 2009-11-11 | 北京中星微电子有限公司 | Control method of traffic signal lamp and system |
| CN101727747A (en) * | 2009-12-16 | 2010-06-09 | 南京信息工程大学 | Abnormal road jam alarming method based on flow detection |
| CN102043945A (en) * | 2010-11-23 | 2011-05-04 | 聊城大学 | License plate character recognition method based on real-time vehicle tracking and binary index classification |
| CN202171874U (en) * | 2011-03-26 | 2012-03-21 | 张奕昕 | Intelligent traffic signal lamp |
| CN102930249A (en) * | 2012-10-23 | 2013-02-13 | 四川农业大学 | Method for identifying and counting farmland pests based on colors and models |
| CN104992223A (en) * | 2015-06-12 | 2015-10-21 | 安徽大学 | Dense population estimation method based on deep learning |
| CN105225500A (en) * | 2015-08-20 | 2016-01-06 | 青岛海信网络科技股份有限公司 | A kind of traffic control aid decision-making method and device |
-
2017
- 2017-08-10 CN CN201710683208.1A patent/CN107689158A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101577054A (en) * | 2009-02-27 | 2009-11-11 | 北京中星微电子有限公司 | Control method of traffic signal lamp and system |
| CN101727747A (en) * | 2009-12-16 | 2010-06-09 | 南京信息工程大学 | Abnormal road jam alarming method based on flow detection |
| CN102043945A (en) * | 2010-11-23 | 2011-05-04 | 聊城大学 | License plate character recognition method based on real-time vehicle tracking and binary index classification |
| CN202171874U (en) * | 2011-03-26 | 2012-03-21 | 张奕昕 | Intelligent traffic signal lamp |
| CN102930249A (en) * | 2012-10-23 | 2013-02-13 | 四川农业大学 | Method for identifying and counting farmland pests based on colors and models |
| CN104992223A (en) * | 2015-06-12 | 2015-10-21 | 安徽大学 | Dense population estimation method based on deep learning |
| CN105225500A (en) * | 2015-08-20 | 2016-01-06 | 青岛海信网络科技股份有限公司 | A kind of traffic control aid decision-making method and device |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108961782A (en) * | 2018-08-21 | 2018-12-07 | 北京深瞐科技有限公司 | Traffic intersection control method and device |
| CN110047272A (en) * | 2019-05-21 | 2019-07-23 | 重庆工程学院 | A kind of intelligent transportation pedestrian behavior monitoring and alarming system based on big data |
| CN110363989A (en) * | 2019-07-11 | 2019-10-22 | 汉王科技股份有限公司 | Magnitude of traffic flow detection method, device, electronic equipment and storage medium |
| CN110826456A (en) * | 2019-10-31 | 2020-02-21 | 青岛海信网络科技股份有限公司 | Countdown board fault detection method and system |
| CN111746728A (en) * | 2020-06-17 | 2020-10-09 | 重庆大学 | Novel overwater cleaning robot based on reinforcement learning and control method |
| CN111746728B (en) * | 2020-06-17 | 2022-06-24 | 重庆大学 | Novel overwater cleaning robot based on reinforcement learning and control method |
| CN112071070A (en) * | 2020-09-18 | 2020-12-11 | 广州维他科技有限公司 | Intelligent traffic intersection controller |
| CN112071070B (en) * | 2020-09-18 | 2021-10-08 | 广州维他科技有限公司 | Intelligent traffic intersection controller |
| CN112991782A (en) * | 2021-04-08 | 2021-06-18 | 河北工业大学 | Control method, system, terminal, equipment, medium and application of traffic signal lamp |
| CN113781784A (en) * | 2021-11-09 | 2021-12-10 | 深圳市奥新科技有限公司 | Intelligent traffic light and control method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107689158A (en) | A kind of intellectual traffic control method based on image procossing | |
| WO2020181685A1 (en) | Vehicle-mounted video target detection method based on deep learning | |
| CN107622678A (en) | A kind of intelligent traffic control system and its method based on image procossing | |
| CN109410577B (en) | Self-adaptive traffic control subarea division method based on space data mining | |
| Fang et al. | Road-sign detection and tracking | |
| CN112001339A (en) | Pedestrian social distance real-time monitoring method based on YOLO v4 | |
| CN106971544B (en) | A kind of direct method that vehicle congestion is detected using still image | |
| CN107067026A (en) | Electrical equipment fault detection method based on deep neural network | |
| CN107689157B (en) | Traffic intersection passable road planning method based on deep learning | |
| CN107273832A (en) | Licence plate recognition method and system based on integrating channel feature and convolutional neural networks | |
| CN110085026A (en) | A kind of traffic status prediction method based on clustering and Markov model | |
| KR102122850B1 (en) | Solution for analysis road and recognition vehicle license plate employing deep-learning | |
| CN108021879A (en) | A kind of vehicular traffic kind identification method based on video image | |
| CN110096981A (en) | A kind of video big data traffic scene analysis method based on deep learning | |
| CN116189099B (en) | Method for detecting and stacking exposed garbage based on improved yolov8 | |
| CN106682681A (en) | Recognition algorithm automatic improvement method based on relevance feedback | |
| Krsák et al. | Traffic sign recognition and localization for databases of traffic signs | |
| Quiros et al. | Machine vision of traffic state estimation using fuzzy logic | |
| CN109993061A (en) | A kind of human face detection and tracing method, system and terminal device | |
| Habermann et al. | Artificial neural nets object recognition for 3D point clouds | |
| CN113095301B (en) | Road occupation operation monitoring method, system and server | |
| CN114926791A (en) | Method and device for detecting abnormal lane change of vehicles at intersection, storage medium and electronic equipment | |
| Wang et al. | Vision-based highway traffic accident detection | |
| Madhu et al. | Identification of different vehicle-following manoeuvres for heterogeneous weak-lane disciplined traffic condition from vehicle trajectory data | |
| CN111178275A (en) | Fire detection method based on convolutional neural network |
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: 20180213 |
|
| RJ01 | Rejection of invention patent application after publication |