CN115116035A - Road traffic light identification system and method based on neural network - Google Patents

Road traffic light identification system and method based on neural network Download PDF

Info

Publication number
CN115116035A
CN115116035A CN202210708301.4A CN202210708301A CN115116035A CN 115116035 A CN115116035 A CN 115116035A CN 202210708301 A CN202210708301 A CN 202210708301A CN 115116035 A CN115116035 A CN 115116035A
Authority
CN
China
Prior art keywords
traffic light
road
target position
neural network
image data
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
Application number
CN202210708301.4A
Other languages
Chinese (zh)
Inventor
朱晓东
刘国清
郑伟
季思文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Youjia Technology Co ltd
Original Assignee
Nanjing Youjia Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Youjia Technology Co ltd filed Critical Nanjing Youjia Technology Co ltd
Priority to CN202210708301.4A priority Critical patent/CN115116035A/en
Publication of CN115116035A publication Critical patent/CN115116035A/en
Pending legal-status Critical Current

Links

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
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/63—Scene text, e.g. street names
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/02—Recognising information on displays, dials, clocks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a road traffic light recognition system and method based on a neural network, belonging to the technical field of intelligent auxiliary driving, and the system comprises: acquiring road image data in front of a vehicle; sending the road image data into a trained traffic light detection neural network framework, and carrying out traffic light target position detection and type identification to obtain traffic light target position information and type identification information; acquiring corresponding traffic light image blocks in the road image data by combining the traffic light target position information and a classified image preprocessing method; and sending the traffic light image blocks into a traffic light identification multi-head network model according to the type identification information to obtain the corresponding traffic light state and countdown information. The invention provides more accurate judgment for the identification of the traffic light countdown numbers so that the intelligent automobile can easily cope with complex environments of intersections.

Description

Road traffic light identification system and method based on neural network
Technical Field
The invention relates to a road traffic light recognition system and method based on a neural network, and belongs to the technical field of intelligent auxiliary driving.
Background
With the development of urban traffic and the progress of the technology level, the intellectualization of automobiles has become the key development direction of each large vehicle enterprise in recent years. Various intelligent devices and technologies are adopted by all the vehicles and enterprises as much as possible to improve the active safety performance of the vehicles and increase the comfort of passengers. Wherein the traffic light identification system has important application value. However, it is difficult to implement a truly practical traffic light identification system for the following reasons: firstly, the traffic light identification method faces an open complex environment, and factors such as variable illumination, size change caused by intersection distance and the like need to be considered; secondly, the traffic light identification method is significant because the application scene must be operated in real time.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a road traffic light identification system and method based on a neural network, which can provide more accurate judgment for the identification of traffic light countdown numbers so that an intelligent automobile can easily cope with complex intersection environments.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the present invention provides a method for identifying a road traffic light based on a neural network, comprising:
acquiring road image data in front of a vehicle;
sending the road image data into a trained traffic light detection neural network framework, and carrying out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
acquiring corresponding traffic light image blocks in the road image data by combining the traffic light target position information and a classified image preprocessing method;
and sending the traffic light image blocks into a traffic light identification multi-head network model according to the type identification information to obtain the corresponding traffic light state and countdown information.
Further, the road image data is acquired by a vehicle-mounted monocular camera mounted on the vehicle in the driving process of the vehicle.
Further, the traffic light detection neural network framework uses a detection framework of yolov5, and the network output of the traffic light detection neural network framework is a layer of feature map, and the traffic light target position information is decoded from the feature map to obtain traffic light target position information and type identification information.
Further, combining the traffic light target position information and a classification image preprocessing method, obtaining a corresponding traffic light image block in the road image data, comprising: in the road image data, a central point and a target frame are set according to traffic light target position information to obtain corresponding traffic light image blocks, wherein coordinates of an upper left point and a lower right point of the traffic light target position information are respectively (x) 1 ,y 1 )、(x 2 ,y 2 ) The center point is
Figure BDA0003706739370000021
Figure BDA0003706739370000022
The width and height of the target frame are all T ═ 1.1 (max (W, H)), W ═ x (x 2 -x 1 ) And H ═ y 2 -y 1 ) And normalizing and scaling the traffic light image blocks to the size required by the classification network.
Furthermore, the traffic light recognition multi-head network model comprises a round lamp body recognition task, an arrow lamp body recognition task and a countdown time recognition task, each task is respectively provided with a training set and a verification set and is respectively sent into a network as different data layers to carry out forward transmission calculation, loss of the corresponding task is obtained in each forward transmission, loss of the three tasks is weighted and summed, then backward transmission iteration network parameters are carried out, and finally the traffic light recognition multi-head network model is obtained.
In a second aspect, the present invention provides a road traffic light recognition system based on a neural network, comprising:
a data acquisition module: for acquiring road image data in front of the vehicle;
traffic light identification module: the system is used for sending road image data into a trained traffic light detection neural network framework to carry out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
an image block acquisition module: the method is used for combining traffic light target position information and a classification image preprocessing method to obtain a corresponding traffic light image block in road image data;
the information state identification module: and the traffic light image blocks are sent to the traffic light recognition multi-head network model according to the type recognition information to obtain the corresponding traffic light state and countdown information.
In a third aspect, the present invention provides a road traffic light recognition device based on a neural network, comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any of the above.
In a fourth aspect, the invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of any of the methods described above.
Compared with the prior art, the invention has the following beneficial effects:
the invention aims to integrate a target detection and target classification method through a neural network technology, firstly, traffic lights are finely divided through a detection method, specific light is identified by using a classification method, meanwhile, some wrong detection results are filtered, and finally, traffic light countdown numbers are identified by using an OCR (optical character recognition) identification mode to realize accurate judgment so as to improve the robustness of the whole system, so that an intelligent automobile can easily cope with complex environments of intersections, and operations such as acceleration, deceleration, start and stop can be accurately finished. The priori scene knowledge is provided, the vehicle is helped to conduct optimization of different safety levels aiming at special scenes, and then more accurate alarm prediction is given.
Drawings
Fig. 1 is a flowchart of a neural network-based road traffic light identification method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a mounting position of a camera according to a first embodiment of the present invention;
FIG. 3 is a schematic diagram of an example of capturing images in a scene presentation according to an embodiment of the present invention;
FIG. 4 is a schematic view of a traffic light body of the type provided in accordance with an embodiment of the present invention;
FIG. 5 is a schematic diagram of a network output modification according to an embodiment of the present invention;
fig. 6 is a schematic diagram of lamp body classification according to an embodiment of the invention;
fig. 7 is a schematic diagram of a network structure of an identification module according to an embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The first embodiment is as follows:
referring to fig. 1, a method for identifying a road traffic light based on a neural network is mainly divided into three parts: acquiring a road image, detecting a road traffic light, and identifying the traffic light. Firstly, acquiring image data of different roads in front of a vehicle in real time through a vehicle-mounted camera, secondly, sending the acquired images into a trained traffic light detection neural network framework to acquire position information and category information of traffic lights in the roads, thirdly, acquiring corresponding Patch according to the vehicle position information and a classification image preprocessing method, and fourthly, sending the Patch into a traffic light identification network according to the category information to acquire the state of the corresponding traffic lights and partial countdown information.
(1) Image acquisition
The vehicle-mounted monocular camera is mounted on the window glass, and vehicle condition information right in front of the vehicle is collected in the running process of the vehicle. The input size of the captured picture is 1920 × 1080P, the position of the camera device is shown in fig. 2, and an example of the captured image in the scene demonstration is shown in fig. 3.
(2) Traffic light detection
Referring to fig. 4, according to various traffic light bodies commonly found on the road surface, the categories are firstly refined in the detection stage, and the light frame and the light body are detected as simultaneously as possible, so as to facilitate the subsequent determination of the relative relationship, and further clarify whether the traffic light bodies are the signals that the vehicle needs to pay attention to. Traffic light detection uses yolov 5's detection frame, adopts MobileNetV2 as the backlight, obtains good detection performance when guaranteeing lightweight network parameter.
Referring to fig. 5, since the distance of the traffic lights is relatively long and the imaging size is relatively small, we modify the Yolov5head network output, and reduce the depth of the network by simplifying the three layers of feature maps with different sizes into a layer of larger feature map, while the single layer of larger feature map focuses on various small-sized targets, and finally decode the corresponding target positions and categories from the feature maps, thereby ensuring that the detection of the traffic lights maintains a relatively high recall rate.
(3) Traffic light identification
Because the traffic lights are distinguished from each other in the traffic light detection, the main bodies and the lamp bodies of different categories need to be classified and identified respectively.
(a) Lamp body classification
Referring to fig. 6, the present solution performs the following classification for the detected traffic light result:
1. a circular lamp body: the color is mainly distinguished, namely red, yellow and green;
2. arrow lamp body: and carrying out refined classification according to colors and different arrow directions.
(b) Number identification
Because different traffic light countdown numbers have different digit problems, the identification work cannot be simply and directly classified, a CTC decoding mode is used, the output of samples is aligned, each digit in the samples is separately classified, and finally the final countdown time is combined.
Pretreatment:
because the sizes of the lamp bodies in the image are different, the lamp bodies can be sent to a classification network to be forwarded to obtain a result after normalization processing is carried out.
The method obtains position information of a lamp body from a traffic light detecting section, wherein (x) 1 ,y1)、(x 2 Y2) represents the coordinates of the upper left point and the lower right point, and the center point is calculated by
Figure BDA0003706739370000061
Width of detection frame W ═ x 2 -x 1 ) The height H ═ y of the detection frame 2 -y 1). Then, the maximum value of the width and the height is expanded by 1.1 times to be used as the width and the height T of a new target frame which is 1.1 (max (W, H)), the corresponding Patch (traffic light image block) is obtained according to the central point P and the target frame T, and finally the Patch is normalized and scaled to the size required by the classification network.
A multi-head network:
in the identification process of the traffic light, the round lamp body, the arrow lamp body and the countdown time identification belong to mutually exclusive tasks, and the tasks are respectively used for model training to cause a large amount of computing resource loss, so that the real-time operation of the system is not facilitated. Therefore, the same network backhaul is used for the three types of tasks, and the mode of multi-head network output is unified into a model.
In order to ensure the precision of each task, a training set and a verification set are required to be respectively set and are respectively sent to a network as different data layers for forward transmission calculation, loss of the corresponding task is obtained in each forward transmission, loss of the three tasks is subjected to weighted summation and then is subjected to backward transmission iteration network parameters, and finally a multi-task comprehensive model is obtained.
Example two:
a road traffic light recognition system based on a neural network, which can implement the road traffic light recognition method based on the neural network according to the first embodiment, includes:
a data acquisition module: for acquiring road image data in front of the vehicle;
traffic light identification module: the traffic light target position detection and type identification system is used for sending road image data into a trained traffic light detection neural network framework to perform traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
an image block acquisition module: the method is used for combining the traffic light target position information and a classified image preprocessing method to obtain corresponding traffic light image blocks in the road image data;
an information state identification module: and the traffic light image blocks are sent to the traffic light recognition multi-head network model according to the type recognition information to obtain the corresponding traffic light state and countdown information.
Example three:
the embodiment of the invention also provides a road traffic light recognition device based on the neural network, which can realize the road traffic light recognition method based on the neural network in the embodiment one and comprises a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method of:
acquiring road image data in front of a vehicle;
sending the road image data into a trained traffic light detection neural network framework, and carrying out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
acquiring corresponding traffic light image blocks in the road image data by combining the traffic light target position information and a classified image preprocessing method;
and sending the traffic light image blocks into a traffic light identification multi-head network model according to the type identification information to obtain the corresponding traffic light state and countdown information.
Example four:
an embodiment of the present invention further provides a computer-readable storage medium, which can implement the method for identifying a road traffic light based on a neural network according to the first embodiment, and a computer program is stored thereon, and when being executed by a processor, the computer program implements the following steps of the method:
acquiring road image data in front of a vehicle;
sending the road image data into a trained traffic light detection neural network framework, and carrying out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
acquiring corresponding traffic light image blocks in the road image data by combining the traffic light target position information and a classified image preprocessing method;
and sending the traffic light image blocks into a traffic light identification multi-head network model according to the type identification information to obtain the corresponding traffic light state and countdown information.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (8)

1. A road traffic light identification method based on a neural network is characterized by comprising the following steps:
acquiring road image data in front of a vehicle;
sending the road image data into a trained traffic light detection neural network framework, and carrying out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
obtaining corresponding traffic light image blocks in the road image data by combining the traffic light target position information and a classified image preprocessing method;
and sending the traffic light image blocks into a traffic light recognition multi-head network model according to the type recognition information to obtain the corresponding traffic light state and countdown information.
2. The neural network-based road traffic light recognition method as claimed in claim 1, wherein the road image data is acquired by a vehicle-mounted monocular camera mounted on a vehicle during a driving process of the vehicle.
3. The neural network-based road traffic light identification method as claimed in claim 1, wherein the traffic light detection neural network framework uses the detection framework of yolov5, and the network output is a layer of feature map, and the traffic light target position information is decoded from the feature map to obtain the traffic light target position information and the type identification information.
4. The neural network-based road traffic light recognition method of claim 1, wherein the method for obtaining corresponding traffic light image blocks in the road image data by combining the traffic light target position information and the classification image preprocessing method comprises: in the road image data, a central point and a target frame are set according to traffic light target position information to obtain corresponding traffic light image blocks, wherein coordinates of an upper left point and a lower right point of the traffic light target position information are respectively (x) 1 ,y 1 )、(x 2 ,y 2 ) The center point is
Figure FDA0003706739360000011
The width and height of the target frame are all T ═ 1.1 (max (W, H)), W ═ x (x 2 -x 1 ) And H ═ y 2 -y 1 ) And normalizing and scaling the traffic light image blocks to the size required by the classification network.
5. The method as claimed in claim 1, wherein the traffic light recognition multi-head network model comprises a round light body recognition task, an arrow light body recognition task and a countdown time recognition task, each task is respectively provided with a training set and a verification set, the training sets and the verification sets are respectively sent into the network as different data layers to perform forwarding calculation, loss of the corresponding task is obtained in each forwarding, loss of the three tasks is weighted and summed, then, backward-transmission iteration network parameters are performed, and finally, the traffic light recognition multi-head network model is obtained.
6. A road traffic light recognition system based on a neural network is characterized by comprising:
a data acquisition module: for acquiring road image data in front of the vehicle;
traffic light identification module: the system is used for sending road image data into a trained traffic light detection neural network framework to carry out traffic light target position detection and type identification to obtain traffic light target position information and type identification information;
an image block acquisition module: the method is used for combining the traffic light target position information and a classified image preprocessing method to obtain corresponding traffic light image blocks in the road image data;
the information state identification module: and the traffic light image blocks are sent to the traffic light recognition multi-head network model according to the type recognition information to obtain the corresponding traffic light state and countdown information.
7. A road traffic light recognition device based on neural network is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 5.
8. Computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 5.
CN202210708301.4A 2022-06-22 2022-06-22 Road traffic light identification system and method based on neural network Pending CN115116035A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210708301.4A CN115116035A (en) 2022-06-22 2022-06-22 Road traffic light identification system and method based on neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210708301.4A CN115116035A (en) 2022-06-22 2022-06-22 Road traffic light identification system and method based on neural network

Publications (1)

Publication Number Publication Date
CN115116035A true CN115116035A (en) 2022-09-27

Family

ID=83329232

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210708301.4A Pending CN115116035A (en) 2022-06-22 2022-06-22 Road traffic light identification system and method based on neural network

Country Status (1)

Country Link
CN (1) CN115116035A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117474905A (en) * 2023-12-26 2024-01-30 广东贝洛新材料科技有限公司 Material property detection method, device, equipment and storage medium
CN117496486A (en) * 2023-12-27 2024-02-02 安徽蔚来智驾科技有限公司 Traffic light shape recognition method, readable storage medium and intelligent device

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117474905A (en) * 2023-12-26 2024-01-30 广东贝洛新材料科技有限公司 Material property detection method, device, equipment and storage medium
CN117496486A (en) * 2023-12-27 2024-02-02 安徽蔚来智驾科技有限公司 Traffic light shape recognition method, readable storage medium and intelligent device
CN117496486B (en) * 2023-12-27 2024-03-26 安徽蔚来智驾科技有限公司 Traffic light shape recognition method, readable storage medium and intelligent device

Similar Documents

Publication Publication Date Title
CN109816024B (en) Real-time vehicle logo detection method based on multi-scale feature fusion and DCNN
CN110069986B (en) Traffic signal lamp identification method and system based on hybrid model
CN105160309B (en) Three lanes detection method based on morphological image segmentation and region growing
De Charette et al. Real time visual traffic lights recognition based on spot light detection and adaptive traffic lights templates
CN115116035A (en) Road traffic light identification system and method based on neural network
CN106971185B (en) License plate positioning method and device based on full convolution network
CN108269407B (en) Security robot capable of automatically managing people stream and logistics
CN106934374B (en) Method and system for identifying traffic signboard in haze scene
CN107506760A (en) Traffic signals detection method and system based on GPS location and visual pattern processing
CN111931683B (en) Image recognition method, device and computer readable storage medium
CN109993138A (en) A kind of car plate detection and recognition methods and device
CN112990065B (en) Vehicle classification detection method based on optimized YOLOv5 model
CN112949578B (en) Vehicle lamp state identification method, device, equipment and storage medium
Riaz et al. YOLO based recognition method for automatic license plate recognition
WO2024051296A1 (en) Method and apparatus for obstacle detection in complex weather
CN111695565A (en) Automobile mark accurate positioning method based on road barrier fuzzy image
CN111860219B (en) High-speed channel occupation judging method and device and electronic equipment
CN112613434A (en) Road target detection method, device and storage medium
CN111814667A (en) Intelligent road condition identification method
CN111444916A (en) License plate positioning and identifying method and system under unconstrained condition
Sung et al. Real-time traffic light recognition on mobile devices with geometry-based filtering
CN111723695A (en) Improved Yolov 3-based driver key sub-area identification and positioning method
CN207115438U (en) Image processing apparatus for vehicle-mounted fisheye camera
CN202771439U (en) Traffic sign automatic identification system based on MATLAB
CN115205651A (en) Low visibility road target detection method based on bimodal fusion

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