CN111950468A - Traffic condition detection method and system based on convolutional neural network and storage medium - Google Patents
Traffic condition detection method and system based on convolutional neural network and storage medium Download PDFInfo
- Publication number
- CN111950468A CN111950468A CN202010818203.7A CN202010818203A CN111950468A CN 111950468 A CN111950468 A CN 111950468A CN 202010818203 A CN202010818203 A CN 202010818203A CN 111950468 A CN111950468 A CN 111950468A
- Authority
- CN
- China
- Prior art keywords
- layer
- neural network
- convolutional neural
- condition detection
- road
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 87
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 66
- 238000011176 pooling Methods 0.000 claims description 35
- 230000004913 activation Effects 0.000 claims description 20
- 238000000034 method Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 9
- 238000011897 real-time detection Methods 0.000 description 6
- 206010039203 Road traffic accident Diseases 0.000 description 5
- 239000002352 surface water Substances 0.000 description 3
- 210000004027 cell Anatomy 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
- G08G1/0112—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
-
- G—PHYSICS
- 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
Abstract
The invention discloses a traffic condition detection method and system based on a convolutional neural network and a storage medium. The traffic condition detection method based on the convolutional neural network comprises the following steps: acquiring a road surface picture acquired by vehicle-mounted equipment; extracting an interested area from the road surface picture, wherein the interested area comprises a road condition characteristic area; inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability. The invention solves the problems of incomplete information collection, instantaneity of detection and the like of the current traffic condition detection method and system, and improves the comprehensiveness and instantaneity of traffic condition detection. In addition, the invention also introduces a plurality of traffic condition classifications, thereby improving the detection precision.
Description
Technical Field
The invention relates to the field of traffic detection, in particular to a traffic condition detection method and system based on a convolutional neural network and a storage medium.
Background
The real-time detection of the road traffic condition is an important reference for a traveler to make a trip plan and is also an important basis for a traffic management department to make management decisions. Real-time detection of road traffic conditions plays an irreplaceable role in intelligent traffic.
However, the existing real-time detection system for road traffic conditions is usually based on a road camera or manual reporting to obtain original information data. However, both of these two methods for acquiring the original information data cannot realize comprehensive and real-time detection of the road traffic condition. On one hand, aiming at manual reporting, because the reporting ways of manual reporting are few, the transmission efficiency is low, and many drivers cannot carry out reporting operation in the driving process, the comprehensiveness and the real-time performance of detection cannot be guaranteed. On the other hand, the original information data is collected aiming at the road camera, and the road camera cannot collect the comprehensive information of the road condition due to scattered distribution points; and the road camera must feed back the image shot by the road camera to the cloud server, and the result can be obtained only by processing data by the cloud server, so that the processing burden of the cloud server is greatly increased, the processing efficiency is reduced, and accurate real-time detection cannot be realized.
Therefore, it is desirable to provide a method for detecting road conditions comprehensively and in real time.
Disclosure of Invention
The invention aims to provide a traffic condition detection method, a system and a storage medium based on a convolutional neural network, so that the traffic condition detection method based on the convolutional neural network can collect and detect the current traffic condition by using a mobile phone program of a driving user, cooperatively collects, processes and feeds back the obtained result to other users through a cloud server, and the comprehensive and real-time detection of the traffic condition is realized.
In order to achieve the purpose, the invention adopts the following technical scheme:
a traffic condition detection method based on a convolutional neural network comprises the following steps:
acquiring a road surface picture acquired by vehicle-mounted equipment;
extracting an interested area from the road surface picture, wherein the interested area comprises a road condition characteristic area;
inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability.
Preferably, in the traffic condition detection method based on the convolutional neural network, the step of extracting the region of interest in the road surface image includes:
and identifying and extracting the region of interest in the road surface picture by an object detection model based on SSDMobileNet.
Preferably, after the step of extracting the region of interest from the road surface image, the method for detecting traffic conditions based on the convolutional neural network further includes:
adjusting the resolution of the region of interest to a target resolution of 256 x 256 pixels.
Preferably, before the step of obtaining the road surface image collected by the vehicle-mounted device, the method for detecting a traffic condition based on a convolutional neural network further includes:
constructing a convolutional neural network, wherein the convolutional neural network comprises an input layer, N activation layers, a full connection layer and an output layer; the activation layer comprises a convolution layer and a pooling layer;
inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result, wherein the step comprises the following steps of:
and a Relu function is adopted in the activation layer as an output activation function of the neuron, after the input value of the input layer passes through the N activation layers, the characteristic data is mapped into a sample space through a full connection layer, and the road condition type and the corresponding probability combination in the region of interest are output and obtained on the output layer.
Preferably, in the traffic condition detection method based on the convolutional neural network, the convolutional neural network includes 3 active layers, which are an active layer one, an active layer two, and an active layer three, respectively;
the first active layer comprises a first convolution layer and a first pooling layer, and an output value of the first convolution layer is used as an input value of the first pooling layer after being activated by a Relu function; the second active layer comprises a second convolution layer and a second pooling layer, and the output value of the second convolution layer is activated by a Relu function and then is used as the input value of the second pooling layer; the third activation layer comprises a third convolution layer and a third pooling layer, and the output value of the third convolution layer is activated by a Relu function and then is used as the input value of the third pooling layer.
Preferably, in the traffic condition detection method based on the convolutional neural network, the scan unit of the convolutional layer one, the convolutional layer two, and the convolutional layer three is a 10 × 10 pixel unit; and the scanning units of the first pooling layer, the second pooling layer and the third pooling layer are 2 x 2 pixel units.
Preferably, in the traffic condition detection method based on the convolutional neural network, after the step of inputting the region of interest into the convolutional neural network trained in advance and obtaining the road condition detection result, the method further includes:
uploading the obtained road condition detection result to a cloud server;
and the cloud server collects the road condition detection result and the corresponding vehicle-mounted equipment positioning, collects, processes and matches the vehicle-mounted equipment to generate an optimal route, and sends the optimal route to the vehicle-mounted equipment.
Preferably, in the traffic condition detection method based on the convolutional neural network, the step of collecting the traffic condition detection result and the corresponding vehicle-mounted device location by the cloud server, performing summary processing and matching with the vehicle-mounted device to generate the preferred route, and sending the preferred route to the vehicle-mounted device includes:
after the cloud server collects the road condition detection results uploaded by all the vehicle users, all the road condition detection results are collected, and then all the road condition detection results are sent to the vehicle-mounted equipment of all the vehicle users; and comparing the current position of each driving user and the destination to which each driving user needs to go according to all road condition detection results to obtain an optimal route through algorithm processing, and sending the recommendation suggestion of the optimal route to the vehicle-mounted equipment of each driving user.
The object of the present invention and the technical problem to be solved by the present invention are also achieved by the following technical means. The traffic condition detection system based on the convolutional neural network is used for realizing any one of the traffic condition detection methods based on the convolutional neural network, and comprises the following steps:
the acquisition unit is used for acquiring a road surface picture acquired by the vehicle-mounted equipment;
the extraction unit is used for extracting an interested area in the road surface picture, and the interested area comprises a road surface area;
the processing unit is used for inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability;
the processing unit comprises a convolutional neural network, wherein the convolutional neural network comprises an input layer, an activation layer, a full connection layer and an output layer; the activation layer includes a convolutional layer and a pooling layer.
The object of the present invention and the technical problem to be solved by the present invention are also achieved by the following technical means. According to the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement any one of the above-mentioned traffic condition detection methods based on a convolutional neural network.
By the technical scheme, the traffic condition detection method, the system and the storage medium based on the convolutional neural network provided by the invention at least have the following advantages:
in the technical scheme of the invention, the traffic condition detection method based on the convolutional neural network can comprehensively feed back the real-time information of the traffic conditions of all roads because a large number of driving users which run on the road surface in real time and are distributed very widely obtain the original information data; meanwhile, the real-time traffic condition of the current position of the driving user is processed and detected by the mobile phone program of the driving user, the workload of the cloud server can be greatly reduced, the cloud server can collect, process and feed back the real-time traffic condition result in a centralized manner more efficiently, an optimal traffic scheme is provided, and the real-time performance of traffic condition detection is improved.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical solutions of the present invention more clearly understood and to implement them in accordance with the contents of the description, the following detailed description is given with reference to the preferred embodiments of the present invention and the accompanying drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive exercise.
Fig. 1 is a flowchart of a traffic condition detection method based on a convolutional neural network according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a convolutional neural network according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a driving user smart phone acquiring a road condition picture according to an embodiment of the present invention.
Fig. 4 is a schematic diagram of a fog picture taken by a smart phone of a driving user according to an embodiment of the present invention.
Fig. 5 is a schematic diagram of identifying an area of interest of the surface water by using an object detection model of SSDMobileNet according to an embodiment of the present invention.
Fig. 6 is a schematic diagram of an algorithm principle of recommending a preferred route by a cloud server in a case provided by the embodiment of the present invention.
In the above figure, a0, fluent; a1, traffic accident; a2, fog.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical scheme of the invention is further explained by the specific implementation mode in combination with the attached drawings.
Example one
An embodiment of the present invention provides a traffic condition detection method based on a convolutional neural network, please refer to fig. 1, where fig. 1 is a flowchart of a traffic condition detection method based on a convolutional neural network provided in an embodiment of the present invention, and specifically includes the following steps:
s1, constructing a convolutional neural network;
specifically, referring to fig. 2, fig. 2 is a schematic structural diagram of a convolutional neural network according to an embodiment of the present invention. The convolutional neural network includes: 1 input layer, 3 convolutional layers, 3 pooling layers, 1 full-link layer, and 1 output layer; the input layer is connected with the first convolution layer, the obtained output is connected with the first pooling layer after being processed by a Relu function, then is connected with the second convolution layer, the obtained output is connected with the second pooling layer after being processed by the Relu function, then is connected with the third convolution layer, the obtained output is connected with the third pooling layer after being processed by the Relu function, then is connected with the full-connection layer, and finally is connected with the output layer to obtain the output;
a first convolution layer, a second convolution layer, a third convolution layer: and carrying out convolution operation on the image, wherein the function of the convolution operation is to scan the image characteristics in blocks and integrate pixels in the image. The scan unit of each convolution layer is a 10 × 10 pixel unit;
a first pooling layer, a second pooling layer and a third pooling layer: and performing pooling operation on the images, acting on feature dimension reduction and image compression, reducing an overfitting phenomenon and improving the fault tolerance of the model. The scan cells of each pooling layer are 2 x 2 pixel cells.
The Relu function is used as an output activation function of a convolutional layer neuron, and the expression of the Relu function is as follows:
f(x)＝max(0，x)
the value obtained by the Relu function is used as an input to the subsequent pooling layer, and the spatial size of the image data can be gradually reduced. The reasonable size of the pooling layer can ensure the classification accuracy and effectively control overfitting, so that the computing resources of the user terminal are reduced.
The specific program parameters of the convolutional layer and the pooling layer are obtained after the training, the adjustment and the testing in the laboratory are finished, the optimized parameters are used as programs issued by driving users, and the optimized parameters can be regularly updated according to the use conditions and the feedback of the users.
S2, acquiring a road surface picture acquired by the vehicle-mounted equipment;
specifically, please refer to fig. 3, and fig. 3 is a schematic diagram illustrating a traffic condition picture collected by a smart phone of a driving user according to an embodiment of the present invention. The vehicle-mounted equipment in the embodiment specifically refers to a smart phone. The specific acquisition method comprises the following steps: the driving user installs the detection program on the smart phone, and the installed detection program can work in real time in the driving process of the driving user. The detection program can shoot the road condition picture of the current position of the driving user in real time by means of the camera of the smart phone of the driving user. Referring to fig. 4, fig. 4 is a schematic view of a fog picture taken by a smart phone of a driving user according to an embodiment of the present invention. The detection program captures the road condition picture as shown in fig. 4 by shooting the road condition picture.
S3, extracting an interested area from the road surface picture;
specifically, the specific implementation manner of step S3 is:
s31: the detection program uses an object detection model of the SSDMobileNet to identify the road condition type characteristic part in the road surface picture of the current position of the driving user. The SSDMobileNet object detection model is an object detection algorithm, and the characteristic part of the road condition type in the road condition picture can be identified and extracted through the object detection model. The extracted area is an area of interest, wherein the area containing the characteristic part of the road condition type is a road condition characteristic area. Referring to fig. 5, fig. 5 is a schematic diagram of an object detection model using SSDMobileNet for identifying a characteristic portion of the surface water according to an embodiment of the present invention. And identifying the characteristic part of the surface water in the image through an SSDMobileNet object detection model, marking the characteristic part by using a square frame, namely an interested area needing to be extracted, wherein the interested area contains a road condition characteristic area.
S32: and the detection program intercepts the identified region of interest from the road condition picture of the current position of the driving user to obtain the characteristic picture of the road condition characteristic part.
S4: adjusting the resolution of the region of interest;
specifically, the detection program compresses the picture of the region of interest into 256 × 256 pixels, and obtains a compressed RGB feature picture as an input of the convolutional neural network.
S5, inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result;
specifically, the region of interest with the resolution adjusted in S4 is used as an input, and the type of traffic condition and the probability of the current position are obtained as an output result through processing by a convolutional neural network constructed in S1 of a detection program installed on the smartphone of the vehicle user;
specifically, after the region of interest is input, after activation of three activation layers, the feature data is mapped into a sample space through a full connection layer, and the type of the feature part and the corresponding correct probability combination are output and obtained on an output layer.
The invention carries out diversified and detailed label division on the detectable traffic condition types, and the detectable traffic condition types comprise: fluency, traffic accidents, heavy fog, road flooding, road surface damage, road construction, and the like. The existing traffic condition detection method can only display traffic jam or smoothness, but cannot carry out detailed explanation on the reason of the traffic jam. The invention can inform the driving user of the specific traffic condition in more detail by refining the detectable traffic condition types, provides a more perfect preferred recommended route for the user and is convenient for the driving user to make more targeted decisions. And the detection program can supplement more detectable traffic condition types in subsequent updating optimization according to the user feedback, so that more detailed traffic condition analysis is provided for the driving users.
S6, each driving user uploads the obtained road condition detection result to a cloud server;
specifically, the specific implementation manner of step S6 is:
and (4) uploading the output results obtained after the step S5 to a cloud server by each vehicle user, and uniformly receiving the output results uploaded by all the vehicle users by the cloud server.
S7, the cloud server collects road condition detection results and corresponding vehicle-mounted equipment positioning, generates an optimal route and sends the optimal route to the vehicle-mounted equipment;
specifically, referring to fig. 6, fig. 6 is a schematic diagram illustrating an algorithm principle that a cloud server sends collected road condition detection results and recommends an optimal route according to an embodiment of the present invention. In the figure, a0 indicates that the road traffic condition detection type is smooth, a1 indicates that the road traffic condition detection type is a traffic accident, and a2 indicates that the road traffic condition detection type is fog. There are A, B, C three lanes from one point of the city to another. When a traffic accident occurs on the road A, so that the lane is crowded, at the moment, an automobile driver on the road A can intelligently recognize that the current road condition is the traffic accident (the character label is A1) through a detection program on the smart phone, and then an output result is uploaded to the cloud server; similarly, the automobile drivers on the road B and the road C recognize that the current road is respectively smooth (character label is a0) and fog (character label is a2) through the detection program on the smart phone, and upload the output result to the cloud server. At the moment, the cloud server can record the road surface conditions of all lanes in real time, collect all output results and send all the output results to all driving users. The server can also recommend a smooth and safe lane to the driving users on the road A and the road C through algorithm processing, so that the road B in the scene is shown.
Example two
Based on the foregoing embodiment, an embodiment of the present invention provides a traffic condition detection system based on a convolutional neural network, which is used to implement the traffic condition detection method based on the convolutional neural network of the foregoing embodiment, and includes:
the acquisition unit is used for acquiring a road surface picture acquired by the vehicle-mounted equipment;
the extraction unit is used for extracting an interested area from the road surface picture, and the interested area comprises a road surface area;
the processing unit is used for inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability;
the processing unit comprises a convolutional neural network, and the convolutional neural network comprises an input layer, an activation layer, a full connection layer and an output layer; the active layer includes a convolutional layer and a pooling layer.
EXAMPLE III
Based on the foregoing embodiments, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements a traffic condition detection method based on a convolutional neural network of the foregoing embodiments.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A traffic condition detection method based on a convolutional neural network is characterized by comprising the following steps:
acquiring a road surface picture acquired by vehicle-mounted equipment;
extracting an interested area from the road surface picture, wherein the interested area comprises a road condition characteristic area;
inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability.
2. The convolutional neural network-based traffic condition detection method as claimed in claim 1, wherein the step of extracting a region of interest in the road surface picture comprises:
and identifying and extracting the region of interest in the road surface picture by an object detection model based on SSDMobileNet.
3. The convolutional neural network-based traffic condition detection method according to claim 1, wherein after extracting the region of interest from the road surface image, the method further comprises:
adjusting the resolution of the region of interest to a target resolution of 256 x 256 pixels.
4. The convolutional neural network-based traffic condition detection method according to claim 1, wherein before acquiring the road surface picture acquired by the vehicle-mounted device, the method further comprises:
constructing a convolutional neural network, wherein the convolutional neural network comprises an input layer, N activation layers, a full connection layer and an output layer; the activation layer comprises a convolution layer and a pooling layer;
inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result, wherein the step comprises the following steps of:
and a Relu function is adopted in the activation layer as an output activation function of the neuron, after the input value of the input layer passes through the N activation layers, the characteristic data is mapped into a sample space through a full connection layer, and the road condition type and the corresponding probability combination in the region of interest are output and obtained on the output layer.
5. The convolutional neural network-based traffic condition detecting method according to claim 4, wherein the convolutional neural network comprises 3 active layers, namely an active layer one, an active layer two and an active layer three;
the first active layer comprises a first convolution layer and a first pooling layer, and an output value of the first convolution layer is used as an input value of the first pooling layer after being activated by a Relu function; the second active layer comprises a second convolution layer and a second pooling layer, and the output value of the second convolution layer is activated by a Relu function and then is used as the input value of the second pooling layer; the third activation layer comprises a third convolution layer and a third pooling layer, and the output value of the third convolution layer is activated by a Relu function and then is used as the input value of the third pooling layer.
6. The convolutional neural network-based traffic condition detecting method of claim 5, wherein the scan unit of convolutional layer one, convolutional layer two, convolutional layer three is a 10 x 10 pixel unit; and the scanning units of the first pooling layer, the second pooling layer and the third pooling layer are 2 x 2 pixel units.
7. The convolutional neural network-based traffic condition detection method as claimed in claim 1, wherein the step of inputting the region of interest into a pre-trained convolutional neural network further comprises, after obtaining the traffic condition detection result:
uploading the obtained road condition detection result to a cloud server;
and the cloud server collects the road condition detection result and the corresponding vehicle-mounted equipment positioning, collects, processes and matches the vehicle-mounted equipment to generate an optimal route, and sends the optimal route to the vehicle-mounted equipment.
8. The convolutional neural network-based traffic condition detection method according to claim 6, wherein the step of collecting the road condition detection result and the corresponding vehicle-mounted device location by the cloud server, performing summary processing and matching with the vehicle-mounted device to generate a preferred route, and sending the preferred route to the vehicle-mounted device comprises:
after the cloud server collects the road condition detection results uploaded by all the vehicle users, all the road condition detection results are collected, and then all the road condition detection results are sent to the vehicle-mounted equipment of all the vehicle users; and comparing the current position of each driving user and the destination to which each driving user needs to go according to all road condition detection results to obtain an optimal route through algorithm processing, and sending the recommendation suggestion of the optimal route to the vehicle-mounted equipment of each driving user.
9. A convolutional neural network-based traffic condition detection system for implementing the convolutional neural network-based traffic condition detection method according to any one of claims 1 to 8, comprising:
the acquisition unit is used for acquiring a road surface picture acquired by the vehicle-mounted equipment;
the extraction unit is used for extracting an interested area in the road surface picture, and the interested area comprises a road surface area;
the processing unit is used for inputting the region of interest into a pre-trained convolutional neural network to obtain a road condition detection result; the road condition detection result comprises a road condition type and a corresponding probability;
the processing unit comprises a convolutional neural network, wherein the convolutional neural network comprises an input layer, an activation layer, a full connection layer and an output layer; the activation layer includes a convolutional layer and a pooling layer.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a convolutional neural network-based traffic condition detection method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010818203.7A CN111950468A (en) | 2020-08-14 | 2020-08-14 | Traffic condition detection method and system based on convolutional neural network and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010818203.7A CN111950468A (en) | 2020-08-14 | 2020-08-14 | Traffic condition detection method and system based on convolutional neural network and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111950468A true CN111950468A (en) | 2020-11-17 |
Family
ID=73342329
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010818203.7A Pending CN111950468A (en) | 2020-08-14 | 2020-08-14 | Traffic condition detection method and system based on convolutional neural network and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111950468A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569769A (en) * | 2019-08-29 | 2019-12-13 | 浙江大搜车软件技术有限公司 | image recognition method and device, computer equipment and storage medium |
| CN111010668A (en) * | 2019-12-16 | 2020-04-14 | 斑马网络技术有限公司 | Information sharing method based on vehicle-mounted terminal position, terminal device and server |
| CN111178237A (en) * | 2019-12-27 | 2020-05-19 | 上海工程技术大学 | Pavement state recognition method |
-
2020
- 2020-08-14 CN CN202010818203.7A patent/CN111950468A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569769A (en) * | 2019-08-29 | 2019-12-13 | 浙江大搜车软件技术有限公司 | image recognition method and device, computer equipment and storage medium |
| CN111010668A (en) * | 2019-12-16 | 2020-04-14 | 斑马网络技术有限公司 | Information sharing method based on vehicle-mounted terminal position, terminal device and server |
| CN111178237A (en) * | 2019-12-27 | 2020-05-19 | 上海工程技术大学 | Pavement state recognition method |
Non-Patent Citations (2)
| Title |
|---|
| 丁松涛;曲仕茹;: "基于深度学习的交通目标感兴趣区域检测", 中国公路学报, no. 09, 15 September 2018 (2018-09-15), pages 171 - 178 * |
| 田锋;雷印杰;邓棋;: "基于YOLOv3的自然路况信息识别研究", 计算机应用研究, no. 1, 30 June 2020 (2020-06-30), pages 401 - 403 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112417953B (en) | Road condition detection and map data updating method, device, system and equipment | |
| CN112380921A (en) | Road detection method based on Internet of vehicles | |
| CN111461083A (en) | Rapid vehicle detection method based on deep learning | |
| WO2022166625A1 (en) | Method for information pushing in vehicle travel scenario, and related apparatus | |
| CN112966665A (en) | Pavement disease detection model training method and device and computer equipment | |
| CN110087041B (en) | Video data processing and transmitting method and system based on 5G base station | |
| CN112767711A (en) | Multi-class multi-scale multi-target snapshot method and system | |
| CN113255605A (en) | Pavement disease detection method and device, terminal equipment and storage medium | |
| CN112562330A (en) | Method and device for evaluating road operation index, electronic equipment and storage medium | |
| CN112749654A (en) | Deep neural network model construction method, system and device for video fog monitoring | |
| CN112464766A (en) | Farmland automatic identification method and system | |
| CN112434566A (en) | Passenger flow statistical method and device, electronic equipment and storage medium | |
| CN112633120A (en) | Intelligent roadside sensing system based on semi-supervised learning and model training method | |
| CN114092920B (en) | Model training method, image classification method, device and storage medium | |
| CN116258940A (en) | Small target detection method for multi-scale features and self-adaptive weights | |
| CN114842285A (en) | Roadside berth number identification method and device | |
| CN113160272B (en) | Target tracking method and device, electronic equipment and storage medium | |
| CN112907972B (en) | Road vehicle flow detection method and system based on unmanned aerial vehicle and computer readable storage medium | |
| CN112597995B (en) | License plate detection model training method, device, equipment and medium | |
| CN112509321A (en) | Unmanned aerial vehicle-based driving control method and system for urban complex traffic situation and readable storage medium | |
| CN113011308A (en) | Pedestrian detection method introducing attention mechanism | |
| CN112288702A (en) | Road image detection method based on Internet of vehicles | |
| CN116229424A (en) | Intelligent networking automobile pedestrian detection method based on improved YOLO v3 | |
| CN111950468A (en) | Traffic condition detection method and system based on convolutional neural network and storage medium | |
| CN114037834B (en) | Semantic segmentation method and device based on fusion of vibration signal and RGB image |
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 |