CN112837541A - Intelligent traffic vehicle flow management method based on improved SSD - Google Patents
Intelligent traffic vehicle flow management method based on improved SSD Download PDFInfo
- Publication number
- CN112837541A CN112837541A CN202011618749.4A CN202011618749A CN112837541A CN 112837541 A CN112837541 A CN 112837541A CN 202011618749 A CN202011618749 A CN 202011618749A CN 112837541 A CN112837541 A CN 112837541A
- Authority
- CN
- China
- Prior art keywords
- traffic flow
- traffic
- junction
- road
- management method
- 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.)
- Granted
Links
Images
Classifications
-
- 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
-
- 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
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0133—Traffic data processing for classifying traffic situation
-
- 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
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to the technical field of intelligent traffic, and particularly discloses an intelligent traffic flow management method based on an improved SSD (solid State disk), which comprises the following steps of S4, acquiring traffic flow condition images through a camera arranged on a detection road section; and step S5, carrying out traffic flow classification analysis on the acquired traffic flow condition images through a traffic flow classification detection model to obtain traffic flow classification detection information. Step S6, analyzing and obtaining the vehicle category of the traffic flow peak time and the maximum traffic flow according to the traffic flow classification detection information; and step S7, when the traffic flow peak time of the road section is detected, the switching and countdown display of the crossing signal lamps of the detected road section and the time length of the traffic signal lamps for displaying the passing of the vehicles are adjusted according to the traffic signal lamp adjusting rule and the traffic flow of the maximum traffic flow. By adopting the technical scheme of the invention, the waiting time of the traffic signal lamp can be acquired and adjusted at the time of traffic flow peak, and the traffic jam condition is relieved.
Description
Technical Field
The invention relates to the technical field of intelligent traffic, in particular to an intelligent traffic vehicle flow management method based on an improved SSD.
Background
The normal operation of the traffic system is to shunt vehicles in the road section to ensure the smoothness of the road section. Traditional traffic systems mainly fall into two categories: firstly, laying and installing a pressure sensor under a specific road section, and sensing the pressure of a vehicle and counting when the vehicle passes by the pressure sensor; and secondly, when a vehicle enters a relevant area, the detector transmits relevant pulses and counts by using ultrasonic waves, infrared rays, radio frequency signals and the like. The first of these systems, when fitted with pressure sensors, destroys the integrity of the ground and, when there are frequent vehicle passes, the life of the system is not very long and the amount of work is very large when it is replaced again; second, the accuracy of the measurement is low when the two vehicles are closer together. In addition, both methods cannot accurately classify the target vehicles, cannot obtain further vehicle information, and only can carry out timed traffic, and have great limitation on the fact that vehicles with crowded peaks at different intersections and different road sections cannot be flexibly time-designed.
With the rapid development of the deep learning theory, the perfection and the continuous improvement of the accuracy based on the detection of computer visual targets and classification counting, a novel intelligent system is produced. The target detection method based on computer vision usually uses a feature extraction method, but the method is labor-consuming, and the target is easy to lose and cannot be connected for detection due to the influence of factors such as the change of environment, the partial deformation of the target, the change of external illumination and the like.
The ssd (single Shot multi box detector) method is a target classification and detection method based on feedforward neural network, which generates a series of target frames with fixed size and probability of class to which the target belongs, and then determines the final detected target by using non-maximum suppression algorithm, but its underlying network VGG-16 occupies about 80% of the calculation time, and thus has a long delay, and is not suitable for real-time detection and classification.
Disclosure of Invention
In order to solve the technical problem that the waiting time of a traffic signal lamp cannot be flexibly adjusted at the time of a traffic flow peak so as to relieve traffic jam, the invention provides an intelligent traffic flow management method based on an improved SSD.
The basic scheme of the invention is as follows:
the intelligent traffic vehicle flow management method based on the improved SSD comprises the following steps:
step S1, designing a basic network;
step S2, collecting picture samples, dividing the picture samples into a training set and a testing set, and initializing the picture samples of the training set;
step S3, under the initial network, training the neural network of the training set, and then using the testing set to evaluate the neural network obtained by training to obtain a traffic flow classification detection model;
step S4, collecting traffic flow condition images through a camera arranged on a detection road section;
and step S5, carrying out traffic flow classification analysis on the acquired traffic flow condition images through a traffic flow classification detection model to obtain traffic flow classification detection information.
Step S6, analyzing according to the traffic flow classification detection information and time to obtain the traffic flow peak time of the detected road section and the vehicle category of the maximum traffic flow;
and step S7, according to the traffic light regulation rule, when the traffic flow peak time of the detected road section is detected, the switching and countdown display of the crossing light of the detected road section is regulated, and the time length of the traffic light for displaying the vehicle passing is regulated according to the vehicle type with the maximum traffic flow.
The basic scheme has the beneficial effects that: 1. according to the technical scheme, the traffic flow classification detection model can be trained by combining a computer technology and a deep learning theory, so that vehicles passing through a detection road section can be distinguished conveniently by different vehicle types, and traffic flow statistics can be carried out respectively according to different types of vehicles.
2. The traffic signal lamps in the detected road section and at the adjacent intersections are switched and displayed in a count-down mode according to the traffic flow classified detection information, so that the condition of a traffic site can be observed in real time, the switching of the traffic signal lamps is controlled in real time, and the condition of traffic jam is relieved.
3. The method comprises the steps that different vehicle types need to pass through intersections in different time periods, if a truck starts slowly, the vehicle body is long, the time for passing through the intersections is longer than that of a car, the time for displaying the vehicle passing through the traffic signal lamp is dynamically adjusted according to the vehicle type of the current maximum vehicle flow passing through the intersections, if the vehicle type of the maximum vehicle flow is the truck, the time for displaying the vehicle passing through the traffic signal lamp is increased, the accessible vehicle flow is increased in the time period for displaying the vehicle passing through the traffic light at a single time, and the traffic jam.
Further, in step S1, the base network adopts a Resenet-50 base network, and the number of convolution kernels from fc6 to conv9 in the base network is reduced to half of the original number.
The beneficial effects are that: the calculation amount of the constructed model is greatly reduced, and meanwhile, the calculation speed is improved.
Further, step S2 specifically includes taking 80% of the collected image samples as a training set, taking 20% of the image samples as a test set, and performing target labeling on the image samples in the training set according to a preset vehicle classification rule.
Has the advantages that: and training the traffic flow classification detection model by using the test set, and evaluating the generalization capability of the traffic flow detection model obtained by training the test set by using the test set.
Further, in step S3, the background picture sample not containing the target in the training set and the picture sample containing the target are trained according to the ratio of 3:1 to obtain a traffic flow classification detection model, and then the picture sample in the testing set is used to evaluate the generalization ability of the traffic flow classification detection model.
Has the advantages that: by training the background picture sample and the picture sample containing the target, the interference of a moving object to the background is reduced.
Further, step S4 includes acquiring a pedestrian flow condition image by a camera disposed on the detection road;
further comprising: step S8, carrying out image analysis on the human flow condition image to obtain the human flow;
and step S9, dynamically adjusting the switching and countdown display of the traffic lights according to the pedestrian flow at the traffic flow peak time according to the preset traffic light adjusting rule.
Has the advantages that: the signal lamps of the intersection are dynamically adjusted according to the pedestrian flow and the vehicle flow of the detected road section, so that the phenomena of road congestion during rush hour on duty and traffic flow and long waiting time of pedestrians during non-rush hour are relieved, and vehicles and pedestrians pass smoothly.
Further, step S8 includes, after performing image analysis on the pedestrian flow condition image, when the obtained pedestrian flow is not zero, timing the pedestrian flow waiting time;
step S9 further includes dynamically adjusting the switching and countdown display of traffic lights according to the traffic light adjustment rule at the traffic flow peak time and the traffic flow waiting time.
Has the advantages that: in the peak time of traffic flow, when the number of people is less, the traffic is mainly taken as the main, thereby avoiding road congestion. And when the flow of people converges to a certain time, or the waiting time of people reaches a certain time, the traffic signal lamp is dynamically adjusted, so that the long-time waiting of people is avoided.
Further, the signal lamp regulation rule is as follows: switching traffic display lamps and displaying the countdown of traffic signals when the traffic flow is in peak time and the pedestrian flow reaches a pedestrian flow threshold; and when the pedestrian flow does not reach the pedestrian flow threshold value and the pedestrian flow waiting time reaches the waiting time threshold value, switching the traffic display lamp and displaying the countdown of the passing traffic signal of the vehicle.
Has the advantages that: and at the peak time of traffic flow, the traffic behavior is taken as the main. However, when a person passes through the intersection, the traffic signal lamps mainly for vehicle passing at present are switched according to the acquired pedestrian flow waiting at the intersection and the waiting time of the person waiting at the intersection, so that the traffic signal lamps at the intersection are dynamically adjusted, the vehicle can pass as far as possible under the condition of ensuring that the person does not block up, and the traffic congestion phenomenon is relieved.
Further, still include:
step S10, respectively collecting the traffic flow condition images of the main road at the junction and the junction through cameras respectively arranged at the junction and the junction;
step S11, the traffic flow classification analysis is carried out on the traffic flow condition images of the main roads of the junction and the driving junction through the traffic flow classification detection model respectively to obtain the traffic flow classification detection information of the main roads of the junction and the driving junction;
step S12, comparing and analyzing the traffic flow classification detection information of the main road of the junction and the main road of the driving junction, if the analysis result is that the main road of the junction is jammed and the main road of the driving junction is not jammed, generating a prompt message according to the driving direction of the driving junction relative to the main road, and sending the prompt message to the driver terminal of the automobile about to pass through the junction;
has the advantages that: if the main road is jammed at the junction of the junction and the main road is not jammed at the next junction of the main road, the jam indicates that a large number of vehicles leave the junction, and therefore the subsequent vehicles entering the junction are shunted by prompting that the subsequent vehicles entering the junction are changed on the road far away from the junction, the main road and the vehicles merging the junction are prevented from being crowded on one lane, and the vehicle jam condition is relieved.
Further, in step S10, when the analysis result shows that the trunk of the merging road is congested and the trunk of the leaving road is not congested, the directions of the leaving road and the merging road relative to the trunk are determined, and the prompt information is generated according to the direction determination result.
Has the advantages that: the same side and the different side of the leaving road and the merging road are used, and the subsequent vehicles entering the road have different lane changing effects on the congestion condition, so that prompt information is generated according to the directions of the leaving road and the merging road relative to the main road, the vehicles are regulated and controlled, and the phenomenon of road congestion is relieved.
Further, if the direction determination result in step S10 is that the directions of the departure road and the entry road are on the same side with respect to the trunk road, the guidance information is generated.
Has the advantages that: when the driving-away road and the converging road are on the same side, the phenomenon of road congestion is aggravated by avoiding the congested road section between the converging road and the driving-away road from changing the road through the prompt information.
Drawings
FIG. 1 is a flow chart of a first embodiment of an SSD-based intelligent transportation traffic flow management method;
FIG. 2 is a test result diagram of a first embodiment of an SSD-based intelligent transportation traffic flow management method;
FIG. 3 is a test result diagram of a first embodiment of an SSD-based intelligent transportation traffic flow management method;
fig. 4 is a flowchart of a second embodiment of the intelligent transportation traffic flow management method based on the improved SSD.
Detailed Description
The following is further detailed by way of specific embodiments:
example one
The intelligent transportation traffic flow management method based on the improved SSD is shown in FIG. 1 and comprises the following steps:
and step S1, designing a basic network. In this embodiment, the base network uses a Resenet-50 base network, and the number of convolution kernels from fc6 to conv9 in the base network is reduced to half of the original number.
And step S2, collecting picture samples, dividing the picture samples into a training set and a testing set, and initializing the picture samples of the training set. The method specifically comprises the steps of taking 80% of collected picture samples as a training set, taking 20% of the picture samples as a testing set, and carrying out target marking on the picture samples in the training set according to a preset vehicle classification rule. The vehicles in this embodiment are classified into six types: cars, taxi cars, vans, buses and motorcycles.
And step S3, carrying out neural network training on the training set under the initial network, and then evaluating the trained neural network by using the test set to obtain a traffic flow classification detection model. Specifically, a background picture sample without a target in the training set and a picture sample with the target are trained according to the proportion of 3:1 to obtain a traffic flow classification detection model, and then the picture sample in the testing set is used for evaluating the generalization capability of the traffic flow classification detection model.
Specifically, in this embodiment, 30000 picture samples are collected and batch-compressed, 80% of the picture samples are selected, and 24000 picture samples are taken as a training set, wherein the objects include 1000 picture samples of cars, taxi, vans, buses, and motorcycles, and the objects are marked, and the marking mode may be manual marking or automatic object marking using an object detection algorithm. And automatically generating a related XML file after the picture sample is subjected to target marking. And (3) training the picture samples in the test set according to the ratio of 3:1 between the background picture sample without the target and the picture sample with the target, and taking the rest 20% of the picture samples as a training set. In the embodiment, the convolutional layer parameters of all picture samples are initialized by using the Msra algorithm, then iteration is performed by using a random gradient descent method (SGD), meanwhile, batch standardization is used for accelerating the convergence rate, the initialized learning rate is 0.1, the attenuation factor is 0.0001, and the impulse is 0.9, and after 138 iterations, training of the traffic flow classification detection model is completed. And (3) testing the traffic flow classification detection model by using the picture samples concentrated in the test, wherein the accuracy of the classification of the target vehicle is more than 98.5%, and the accuracy of the detection of the target vehicle reaches 99.2%.
And step S4, acquiring traffic flow condition images through the cameras arranged on the detection road sections. In this embodiment, the camera is a high-definition camera that collects in real time and has a resolution of 1280 × 720, and the resolution of the collected picture is too high, which may slow down the detection speed of the system, and in this embodiment, the collected traffic flow status image is further compressed into 420 × 300 pictures.
And step S5, carrying out traffic flow classification analysis on the acquired traffic flow condition images through a traffic flow classification detection model to obtain traffic flow classification detection information. Specifically, a detection frame is arranged at a fixed position of the traffic flow condition image, when a target touches the detection frame, the number of classifiers is increased by 1, and the traffic flow classification detection model classifies the target vehicle type. The test results are shown in fig. 2 and 3.
And step S6, analyzing according to the traffic flow classification detection information and time to obtain the traffic flow peak time of the detected road section and the vehicle category of the maximum traffic flow. Different types of vehicles have different lengths, so that the maximum number of vehicles which can be accommodated in the road section can be calculated according to the traffic flow classification detection information, namely whether the road section is congested or not is judged.
And step S7, according to the traffic light regulation rule, when the traffic flow peak time of the detected road section is detected, the switching and countdown display of the crossing light of the detected road section is regulated, and the time length of the traffic light for displaying the vehicle passing is regulated according to the vehicle type with the maximum traffic flow. When the detected road section is at the time of traffic flow peak, the intersection signal lamp is switched to be a long-time vehicle passing state, and the countdown of people passing the road is displayed.
Example two
The difference from the first embodiment is that: as shown in fig. 4, the method further includes:
in step S4, the method further includes collecting a pedestrian flow condition image by a camera disposed on the detection road.
And step S8, carrying out image analysis on the human flow condition image to obtain the human flow. In this embodiment, step S8 further includes, after performing image analysis on the image of the pedestrian flow condition, when the obtained pedestrian volume is not zero, timing the pedestrian flow waiting duration;
and step S9, dynamically adjusting the switching and countdown display of the traffic lights according to the pedestrian flow at the traffic flow peak time according to the preset traffic light adjusting rule. In order to avoid the situation that the traffic jam is not reached, but the waiting time of the traffic flow waiting for passing is too long, step S9 further includes dynamically adjusting the switching and countdown display of the traffic light according to the traffic light adjustment rule and the traffic flow and the waiting time of the traffic flow at the traffic flow peak time. In this embodiment, the signal lamp adjustment rule is: switching traffic display lamps and displaying the countdown of traffic signals when the traffic flow is in peak time and the pedestrian flow reaches a pedestrian flow threshold; and when the pedestrian flow does not reach the pedestrian flow threshold value and the pedestrian flow waiting time reaches the waiting time threshold value, switching the traffic display lamp and displaying the countdown of the passing traffic signal of the vehicle.
EXAMPLE III
The difference from the first embodiment is that: further comprising step S9;
step S10, respectively collecting the traffic flow condition images of the main road at the junction and the junction through cameras respectively arranged at the junction and the junction;
step S11, the traffic flow classification analysis is carried out on the traffic flow condition images of the main roads of the junction and the driving junction through the traffic flow classification detection model respectively to obtain the traffic flow classification detection information of the main roads of the junction and the driving junction;
and step S12, comparing and analyzing the traffic flow classification detection information of the main road of the junction and the main road of the driving intersection, if the analysis result is that the main road of the junction is jammed and the main road of the driving intersection is not jammed, generating prompt information according to the driving direction of the driving intersection compared with the main road, and sending the prompt information to the driver terminal of the automobile about to pass through the junction, and particularly positioning the driver terminal of the automobile according to the positioning information of the driver.
When the vehicles leave the same side and the different sides of the merging road, the subsequent vehicles enter the same side, and the effects of lane changing on congestion conditions are inconsistent, and when the vehicles leave the same side of the merging road, the congested road section between the merging road and the leaving road is prevented from changing lanes through lane changing, so that the road congestion phenomenon is aggravated. Therefore, in this embodiment, step S10 further includes, when the analysis result indicates that the trunk of the merging road is congested and the trunk of the leaving road is not congested, performing direction determination on the leaving road and the merging road relative to the trunk, and generating the prompt information according to the direction determination result. Specifically, if the direction judgment result is that the directions of the departure road and the entry road relative to the trunk road are the same, prompt information is generated.
The foregoing is merely an example of the present invention and common general knowledge of known specific structures and features of the embodiments is not described herein in any greater detail. It should be noted that, for those skilled in the art, without departing from the structure of the present invention, several changes and modifications can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicability of the patent. The scope of the claims of the present application shall be determined by the contents of the claims, and the description of the embodiments and the like in the specification shall be used to explain the contents of the claims.
Claims (10)
1. The intelligent traffic vehicle flow management method based on the improved SSD is characterized by comprising the following steps of:
step S1, designing a basic network;
step S2, collecting picture samples, dividing the picture samples into a training set and a testing set, and initializing the picture samples of the training set;
step S3, under the initial network, training the neural network of the training set, and then using the testing set to evaluate the neural network obtained by training to obtain a traffic flow classification detection model;
step S4, collecting traffic flow condition images through a camera arranged on a detection road section;
and step S5, carrying out traffic flow classification analysis on the acquired traffic flow condition images through a traffic flow classification detection model to obtain traffic flow classification detection information.
Step S6, analyzing according to the traffic flow classification detection information and time to obtain the traffic flow peak time of the detected road section and the vehicle category of the maximum traffic flow;
and step S7, according to the traffic light regulation rule, when the traffic flow peak time of the detected road section is detected, the switching and countdown display of the crossing light of the detected road section is regulated, and the time length of the traffic light for displaying the vehicle passing is regulated according to the vehicle type with the maximum traffic flow.
2. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 1, wherein: in step S1, the base network uses a Resenet-50 base network, and the number of convolution kernels from fc6 to conv9 in the base network is reduced to half of the original number.
3. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 1, wherein: in step S2, the method specifically includes taking 80% of the collected picture samples as a training set, taking 20% of the picture samples as a test set, and performing target labeling on the picture samples in the training set according to a preset vehicle classification rule.
4. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 3, wherein: in step S3, the background picture samples that do not contain the target in the training set and the picture samples that contain the target are trained according to a ratio of 3:1 to obtain a traffic flow classification detection model, and then the picture samples in the testing set are used to evaluate the generalization ability of the traffic flow classification detection model.
5. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 1, wherein: step S4, collecting people stream status images through a camera arranged on the detection road section;
further comprising: step S8, carrying out image analysis on the human flow condition image to obtain the human flow;
and step S9, dynamically adjusting the switching and countdown display of the traffic lights according to the pedestrian flow at the traffic flow peak time according to the preset traffic light adjusting rule.
6. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 5, wherein: step S8 further comprises the steps of timing the pedestrian flow waiting time when the obtained pedestrian flow is not zero after image analysis is carried out on the pedestrian flow condition images;
step S9 further includes dynamically adjusting the switching and countdown display of traffic lights according to the traffic light adjustment rule at the traffic flow peak time and the traffic flow waiting time.
7. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 5, wherein: the signal lamp adjusting rule is as follows: switching traffic display lamps when the traffic flow reaches a pedestrian flow threshold value at the peak time of the traffic flow, and displaying the countdown of traffic signals when the vehicles pass through; and when the pedestrian flow does not reach the pedestrian flow threshold value and the pedestrian flow waiting time reaches the waiting time threshold value, switching the traffic display lamp and displaying the countdown of the passing traffic signal of the vehicle.
8. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 1, wherein: further comprising:
step S10, respectively collecting the traffic flow condition images of the main road at the junction and the junction through cameras respectively arranged at the junction and the junction;
step S11, the traffic flow classification analysis is carried out on the traffic flow condition images of the main roads of the junction and the driving junction through the traffic flow classification detection model respectively to obtain the traffic flow classification detection information of the main roads of the junction and the driving junction;
and step S12, comparing and analyzing the traffic flow classification detection information of the main road of the junction and the main road of the driving junction, if the analysis result is that the main road of the junction is jammed and the main road of the driving junction is not jammed, generating prompt information according to the driving direction of the driving junction relative to the main road, and sending the prompt information to the driver terminal of the automobile about to pass through the junction.
9. The intelligent transportation traffic flow management method based on the improved SSD as set forth in claim 8, wherein: and step S10, when the analysis result shows that the main road of the junction is jammed and the main road of the junction is not jammed, the orientation of the main road relative to the departure road and the junction is judged, and prompt information is generated according to the orientation judgment result.
10. The intelligent transportation traffic flow management method based on the improved SSD according to claim 9, characterized in that: if the direction determination result in step S10 is that the directions of the departure road and the entry road with respect to the trunk road are the same, the guidance information is generated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011618749.4A CN112837541B (en) | 2020-12-31 | 2020-12-31 | Intelligent traffic vehicle flow management method based on improved SSD |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011618749.4A CN112837541B (en) | 2020-12-31 | 2020-12-31 | Intelligent traffic vehicle flow management method based on improved SSD |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112837541A true CN112837541A (en) | 2021-05-25 |
| CN112837541B CN112837541B (en) | 2022-04-29 |
Family
ID=75925698
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011618749.4A Active CN112837541B (en) | 2020-12-31 | 2020-12-31 | Intelligent traffic vehicle flow management method based on improved SSD |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112837541B (en) |
Citations (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202584432U (en) * | 2012-05-25 | 2012-12-05 | 北京易华录信息技术股份有限公司 | Smart ramp control system |
| JP2014052341A (en) * | 2012-09-10 | 2014-03-20 | Hitachi Automotive Systems Ltd | Map creation server, map creation method and map creation system |
| CN106601002A (en) * | 2016-11-23 | 2017-04-26 | 苏州大学 | City expressway access ramp vehicle pass guiding system in car networking environment and guiding method thereof |
| CN107016861A (en) * | 2017-05-31 | 2017-08-04 | 电子科技大学 | Traffic lights intelligent control system based on deep learning and intelligent road-lamp |
| CN107134156A (en) * | 2017-06-16 | 2017-09-05 | 上海集成电路研发中心有限公司 | A kind of method of intelligent traffic light system and its control traffic lights based on deep learning |
| CN107316468A (en) * | 2017-06-14 | 2017-11-03 | 苏州远征魂车船技术有限公司 | A kind of road intelligent control device |
| CN108960015A (en) * | 2017-05-24 | 2018-12-07 | 优信拍（北京）信息科技有限公司 | A kind of vehicle system automatic identifying method and device based on deep learning |
| US20190120637A1 (en) * | 2017-10-25 | 2019-04-25 | Tata Consultancy Services Limited | Predicting vehicle travel time on routes of unbounded length in arterial roads |
| KR20190048979A (en) * | 2017-10-31 | 2019-05-09 | (주)노바코스 | Vehicle classification system using non-contact automatic vehicle detectior |
| CN109754621A (en) * | 2019-03-01 | 2019-05-14 | 苏州星奥达科技有限公司 | A kind of video analysis method based on Freeway Conditions situation |
| CN110033479A (en) * | 2019-04-15 | 2019-07-19 | 四川九洲视讯科技有限责任公司 | Traffic flow parameter real-time detection method based on Traffic Surveillance Video |
| CN110334572A (en) * | 2019-04-04 | 2019-10-15 | 南京航空航天大学 | The fine recognition methods of vehicle under a kind of multi-angle |
| CN110570670A (en) * | 2019-10-16 | 2019-12-13 | 公安部交通管理科学研究所 | Road section pedestrian crossing signal system and method based on video pedestrian detection |
| CN110633751A (en) * | 2019-09-17 | 2019-12-31 | 上海眼控科技股份有限公司 | Training method of car logo classification model, car logo identification method, device and equipment |
| CN110706482A (en) * | 2019-08-30 | 2020-01-17 | 安锐 | Peak traffic passing control method and device |
| CN111126472A (en) * | 2019-12-18 | 2020-05-08 | 南京信息工程大学 | Improved target detection method based on SSD |
| CN111127908A (en) * | 2019-12-30 | 2020-05-08 | 戴姆勒股份公司 | Automatic driving control method for vehicle confluence |
| CN111369813A (en) * | 2020-03-23 | 2020-07-03 | 江苏大学 | Ramp division and confluence cooperative control method and system for intelligent network-connected automobile |
| CN111523482A (en) * | 2020-04-24 | 2020-08-11 | 深圳市商汤科技有限公司 | Lane congestion detection method and apparatus, electronic device, and storage medium |
| CN111583674A (en) * | 2020-03-23 | 2020-08-25 | 南京慧尔视智能科技有限公司 | Interleaved road section and method for improving traffic efficiency thereof through radar tracking control |
| CN111613072A (en) * | 2020-05-08 | 2020-09-01 | 上海数道信息科技有限公司 | Intelligent signal lamp timing optimization method, device, equipment, system and medium |
| CN111627228A (en) * | 2020-04-03 | 2020-09-04 | 北京博研智通科技有限公司 | Expressway confluence point lane lamp control system and method |
| CN111739312A (en) * | 2020-06-11 | 2020-10-02 | 湖北美和易思教育科技有限公司 | Artificial intelligence thing networking traffic lights system |
| CN111783685A (en) * | 2020-05-08 | 2020-10-16 | 西安建筑科技大学 | Target detection improved algorithm based on single-stage network model |
| CN111785022A (en) * | 2020-07-13 | 2020-10-16 | 上海市城市建设设计研究总院（集团）有限公司 | Urban expressway main and auxiliary system traffic arrangement structure and control method |
| CN111860499A (en) * | 2020-07-01 | 2020-10-30 | 电子科技大学 | Feature grouping-based bilinear convolutional neural network automobile brand identification method |
| CN111862632A (en) * | 2020-04-29 | 2020-10-30 | 南阳师范学院 | Method and device for preventing congestion of intelligent traffic signal lamp |
| CN111899530A (en) * | 2020-06-22 | 2020-11-06 | 厦门快商通科技股份有限公司 | Traffic light control method, system, mobile terminal and storage medium |
| CN112150835A (en) * | 2020-09-07 | 2020-12-29 | 高桂阳 | Urban intelligent traffic management system |
-
2020
- 2020-12-31 CN CN202011618749.4A patent/CN112837541B/en active Active
Patent Citations (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202584432U (en) * | 2012-05-25 | 2012-12-05 | 北京易华录信息技术股份有限公司 | Smart ramp control system |
| JP2014052341A (en) * | 2012-09-10 | 2014-03-20 | Hitachi Automotive Systems Ltd | Map creation server, map creation method and map creation system |
| CN106601002A (en) * | 2016-11-23 | 2017-04-26 | 苏州大学 | City expressway access ramp vehicle pass guiding system in car networking environment and guiding method thereof |
| CN108960015A (en) * | 2017-05-24 | 2018-12-07 | 优信拍（北京）信息科技有限公司 | A kind of vehicle system automatic identifying method and device based on deep learning |
| CN107016861A (en) * | 2017-05-31 | 2017-08-04 | 电子科技大学 | Traffic lights intelligent control system based on deep learning and intelligent road-lamp |
| CN107316468A (en) * | 2017-06-14 | 2017-11-03 | 苏州远征魂车船技术有限公司 | A kind of road intelligent control device |
| CN107134156A (en) * | 2017-06-16 | 2017-09-05 | 上海集成电路研发中心有限公司 | A kind of method of intelligent traffic light system and its control traffic lights based on deep learning |
| US20190120637A1 (en) * | 2017-10-25 | 2019-04-25 | Tata Consultancy Services Limited | Predicting vehicle travel time on routes of unbounded length in arterial roads |
| KR20190048979A (en) * | 2017-10-31 | 2019-05-09 | (주)노바코스 | Vehicle classification system using non-contact automatic vehicle detectior |
| CN109754621A (en) * | 2019-03-01 | 2019-05-14 | 苏州星奥达科技有限公司 | A kind of video analysis method based on Freeway Conditions situation |
| CN110334572A (en) * | 2019-04-04 | 2019-10-15 | 南京航空航天大学 | The fine recognition methods of vehicle under a kind of multi-angle |
| CN110033479A (en) * | 2019-04-15 | 2019-07-19 | 四川九洲视讯科技有限责任公司 | Traffic flow parameter real-time detection method based on Traffic Surveillance Video |
| CN110706482A (en) * | 2019-08-30 | 2020-01-17 | 安锐 | Peak traffic passing control method and device |
| CN110633751A (en) * | 2019-09-17 | 2019-12-31 | 上海眼控科技股份有限公司 | Training method of car logo classification model, car logo identification method, device and equipment |
| CN110570670A (en) * | 2019-10-16 | 2019-12-13 | 公安部交通管理科学研究所 | Road section pedestrian crossing signal system and method based on video pedestrian detection |
| CN111126472A (en) * | 2019-12-18 | 2020-05-08 | 南京信息工程大学 | Improved target detection method based on SSD |
| CN111127908A (en) * | 2019-12-30 | 2020-05-08 | 戴姆勒股份公司 | Automatic driving control method for vehicle confluence |
| CN111583674A (en) * | 2020-03-23 | 2020-08-25 | 南京慧尔视智能科技有限公司 | Interleaved road section and method for improving traffic efficiency thereof through radar tracking control |
| CN111369813A (en) * | 2020-03-23 | 2020-07-03 | 江苏大学 | Ramp division and confluence cooperative control method and system for intelligent network-connected automobile |
| CN111627228A (en) * | 2020-04-03 | 2020-09-04 | 北京博研智通科技有限公司 | Expressway confluence point lane lamp control system and method |
| CN111523482A (en) * | 2020-04-24 | 2020-08-11 | 深圳市商汤科技有限公司 | Lane congestion detection method and apparatus, electronic device, and storage medium |
| CN111862632A (en) * | 2020-04-29 | 2020-10-30 | 南阳师范学院 | Method and device for preventing congestion of intelligent traffic signal lamp |
| CN111613072A (en) * | 2020-05-08 | 2020-09-01 | 上海数道信息科技有限公司 | Intelligent signal lamp timing optimization method, device, equipment, system and medium |
| CN111783685A (en) * | 2020-05-08 | 2020-10-16 | 西安建筑科技大学 | Target detection improved algorithm based on single-stage network model |
| CN111739312A (en) * | 2020-06-11 | 2020-10-02 | 湖北美和易思教育科技有限公司 | Artificial intelligence thing networking traffic lights system |
| CN111899530A (en) * | 2020-06-22 | 2020-11-06 | 厦门快商通科技股份有限公司 | Traffic light control method, system, mobile terminal and storage medium |
| CN111860499A (en) * | 2020-07-01 | 2020-10-30 | 电子科技大学 | Feature grouping-based bilinear convolutional neural network automobile brand identification method |
| CN111785022A (en) * | 2020-07-13 | 2020-10-16 | 上海市城市建设设计研究总院（集团）有限公司 | Urban expressway main and auxiliary system traffic arrangement structure and control method |
| CN112150835A (en) * | 2020-09-07 | 2020-12-29 | 高桂阳 | Urban intelligent traffic management system |
Non-Patent Citations (4)
| Title |
|---|
| GÜLYETER ÖZTÜRK: "Recognition of Vehicles, Pedestrians and Traffic Signs Using Convolutional Neural Networks", 《2020 4TH INTERNATIONAL SYMPOSIUM ON MULTIDISCIPLINARY STUDIES AND INNOVATIVE TECHNOLOGIES (ISMSIT)》 * |
| 敖鸿飞: "基于深度学习的影视图片的服装匹配系统", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
| 李小宁: "基于改进SSD 的车辆小目标检测方法", 《应用光学》 * |
| 熊中刚，贺娟，敖邦乾，叶振环，令狐金卿，罗素莲: "基于路径弯曲度动态预瞄搜索算法的车辆路径跟踪控制软件系统设计", 《江苏农业科学》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112837541B (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111368687B (en) | Sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation | |
| CN106023605B (en) | A kind of method for controlling traffic signal lights based on depth convolutional neural networks | |
| US9418546B1 (en) | Traffic detection with multiple outputs depending on type of object detected | |
| EP1930863B1 (en) | Detecting and recognizing traffic signs | |
| CN110738857B (en) | Vehicle violation evidence obtaining method, device and equipment | |
| CN110287905B (en) | Deep learning-based real-time traffic jam area detection method | |
| CN110660234B (en) | Intelligent traffic light control method | |
| CN105512623A (en) | Foggy-day driving visual enhancement and visibility early warning system and method based on multiple sensors | |
| CN112017437B (en) | Intersection traffic information perception control system and method | |
| US20030190058A1 (en) | Apparatus and method for measuring queue length of vehicles | |
| CN109272482B (en) | Urban intersection vehicle queuing detection system based on sequence images | |
| CN113345237A (en) | Lane-changing identification and prediction method, system, equipment and storage medium for extracting vehicle track by using roadside laser radar data | |
| CN110598511A (en) | Method, device, electronic equipment and system for detecting green light running event | |
| Chebrolu et al. | Deep learning based pedestrian detection at all light conditions | |
| US20210053572A1 (en) | Vehicular driving assist system with traffic jam probability determination | |
| CN106372619A (en) | Vehicle robustness detection and divided-lane arrival accumulative curve estimation method | |
| Bharadwaj et al. | Traffic data analysis using image processing technique on Delhi–Gurgaon expressway | |
| KR20080004833A (en) | Apparatus and method for detecting a navigation vehicle in day and night according to luminous state | |
| Hung et al. | A traffic monitoring system for a mixed traffic flow via road estimation and analysis | |
| CN112837541B (en) | Intelligent traffic vehicle flow management method based on improved SSD | |
| Zheng et al. | Detecting cycle failures at signalized intersections using video image processing | |
| CN1573797A (en) | Method and apparatus for improving the identification and/or re-identification of objects in image processing | |
| Zhang et al. | Machine learning and computer vision-enabled traffic sensing data analysis and quality enhancement | |
| Špaňhel et al. | Detection of traffic violations of road users based on convolutional neural networks | |
| CN112907978A (en) | Traffic flow monitoring method based on monitoring video |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |