CN115147869A - Electric vehicle rider helmet detection and license plate recognition method based on deep learning - Google Patents
Electric vehicle rider helmet detection and license plate recognition method based on deep learning Download PDFInfo
- Publication number
- CN115147869A CN115147869A CN202210750226.8A CN202210750226A CN115147869A CN 115147869 A CN115147869 A CN 115147869A CN 202210750226 A CN202210750226 A CN 202210750226A CN 115147869 A CN115147869 A CN 115147869A
- Authority
- CN
- China
- Prior art keywords
- license plate
- electric vehicle
- detection
- model
- helmet
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- 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
- G06N3/08—Learning methods
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- G06V20/625—License plates
Abstract
The invention discloses a deep learning-based electric vehicle rider helmet detection and license plate recognition method, which belongs to the technical field of image and video intelligent analysis and comprises a self-built electric vehicle rider helmet and a license plate detection data set; constructing a detection model based on improved YOLOv5m, and training the model by using an electric vehicle rider helmet and a license plate detection data set; carrying out segmentation correction on the detected electric vehicle license plate, and carrying out data enhancement on the segmented and corrected license plate picture to establish a license plate identification data set; constructing an electric vehicle license plate recognition model based on the improved CRNN, and training the model by using a license plate recognition data set; and inputting the real-time traffic intersection video into the trained model, detecting the wearing condition of the helmet of the rider, displaying the license plate number of the electric vehicle without the helmet of the rider and the like. The invention solves the problems of low accuracy and low running speed of the traditional detection and identification method, reduces the missing rate of the model under the condition of multiple targets, and improves the identification accuracy of the electric vehicle license plate.
Description
Technical Field
The invention relates to the technical field of intelligent analysis of images and videos, in particular to a method for detecting a rider helmet of an electric vehicle and identifying a license plate of the electric vehicle based on deep learning.
Background
With the development of science and technology and the improvement of the living standard of people, more and more people select the electric vehicle as a vehicle when going out, but with the increase of electric vehicles, various illegal problems of the electric vehicle emerge endlessly, the problem that the electric vehicle is not worn by a helmet when being ridden is particularly prominent, and the working pressure of traffic management departments is directly increased. According to the statistical analysis of the traffic administration of the ministry of public security, 80.0% of the accidents caused by the death of the drivers and the working personnel of the electric bicycle are caused by the injury of the cranium and the brain. Some related research results show that the helmet can reduce the average death rate and the risk action range of various road traffic safety accidents by 60.0-70.0% by accurately wearing the helmet. The safety helmet is a necessary measure for ensuring safe travel.
At present, most methods for detecting and identifying the electric vehicle are used for independently detecting the rider helmet of the electric vehicle or identifying the license plate of the electric vehicle, and meanwhile, the methods for detecting the rider helmet and identifying the license plate of the electric vehicle are few, but the traditional detection method has the problems of low detection accuracy, low detection speed and the like. The license plate of the electric vehicle is mostly recognized by adopting a method of dividing the license plate, and the method can reduce the accuracy of license plate recognition. Most helmet detection and license plate recognition methods detect and recognize collected pictures, but not collected videos.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a method for detecting a helmet of a rider of an electric vehicle and identifying a license plate of the electric vehicle based on deep learning, which can realize the detection of the wearing condition of the helmet of the rider of the electric vehicle and the identification of the license plate of the electric vehicle, and improve the speed and the precision of the helmet detection of the rider of the electric vehicle and the identification of the license plate.
The purpose of the invention can be realized by the following technical scheme:
a deep learning-based electric vehicle rider helmet detection and license plate recognition method comprises the following steps:
step one, self-building an electric vehicle rider helmet and a license plate detection data set;
step two, constructing a detection model based on improved YOLOv5m, and training the model by using the electric vehicle rider helmet and the license plate detection data set;
thirdly, segmenting and correcting the electric vehicle license plate in the detection result, and performing data enhancement on the segmented and corrected license plate picture to establish a license plate identification data set;
constructing an electric vehicle license plate recognition model based on the improved CRNN, and training the model by using a license plate recognition data set;
and fifthly, inputting the real-time traffic intersection video into the trained model, detecting the condition that the rider does not wear the helmet in the video, and displaying the license plate number of the electric vehicle without the helmet.
Further, the data collection in the first step selects the traffic environment of the traffic intersection as the research background, based on spring, summer, autumn and winter, the scene of selecting the different clothes of the passerby is shot the collection to the electric motor car picture of riding.
Furthermore, in the first step, aiming at the problems that the shot data set is insufficient and has unbalanced samples, the shot data set needs to be expanded through data enhancement algorithms such as increasing picture brightness, randomly rotating, reducing picture brightness, randomly cutting, adding noise and the like to obtain a detection data set of a driver helmet and a license plate of the self-built electric vehicle, the data set determines 2-3 detection targets such as the helmet and the license plate and the like, and open source software labelImg is adopted as a training set sample labeling tool.
Furthermore, in the second step, a CBAM attention module is added to the Neck part of the YOLOv5m model for detecting the small target and the medium target, so that the YOLOv5m model pays attention to the channel characteristics of the information of the small target, thereby enhancing the detection capability of the YOLOv5m model for detecting the small target.
Furthermore, the GIoULoss is replaced by the CIoULoss, so that the CIoULoss is used as a regression loss function of the target boundary box, the regression stability of the target boundary box is enhanced, the accuracy of target prediction is improved, the cascade NMS is replaced by the DIoU-NMS, and the undetected rate under the condition of target crowding is reduced.
Further, in the third step, the license plate picture is detected through an improved YOLOv5m detection model, then the inclined license plate of the electric vehicle is corrected through affine transformation, and the number of the license plates of the electric vehicle needs to be expanded through data enhancement algorithms such as increasing the picture brightness, randomly rotating, reducing the picture brightness, randomly cutting, adding noise and the like.
Furthermore, the step four of constructing the improved CRNN-based electric vehicle license plate recognition model is to fuse a CTPN character detection model and the CRNN character recognition model together, carry out CTPN character detection before carrying out CRNN character recognition, respectively detect two lines or one line of characters of the electric vehicle license plate by the CTPN character detection model, and send the two lines or one line of characters into the CRNN electric vehicle license plate recognition model.
Further, in the fifth step, the wearing condition of the helmet of the rider of the electric vehicle in the video is detected based on an improved YOLOv5m detection model, if the rider does not wear the helmet and the electric vehicle has a license plate, the head of the rider and the license plate of the electric vehicle in the video are marked, the detection result based on the improved YOLOv5m detection model is sent to an improved CRNN-based electric vehicle license plate recognition model, the input license plate is recognized, and the recognized license plate number is displayed in the video.
The invention has the beneficial effects that:
1. the invention identifies the license plate number of the electric vehicle while detecting that the electric vehicle rider does not wear the helmet, can help the traffic control department to find the person who does not wear the helmet, and reduces the working pressure of the traffic control department; meanwhile, the invention can directly detect and identify the real-time traffic intersection video and display the detection result in the video, thereby supervising and urging the electric vehicle rider to wear the helmet and ensuring the life safety of the electric vehicle rider.
2. According to the method, the number of pictures is increased through a data enhancement method, the shooting environment is enriched, the model has better generalization, the detection accuracy and the detection speed are improved based on the improved YOLOv5m detection model compared with the traditional method, and the target omission ratio under the condition of target crowding is reduced.
3. The electric vehicle license plate recognition model based on the improved CRNN is constructed, the problem that two lines or one line of characters of the electric vehicle license plate are inconvenient to divide is solved, the electric vehicle license plate recognition can be carried out without division, and the accuracy of license plate recognition is greatly improved.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a flow chart of a deep learning-based method for detecting a rider's helmet and recognizing a license plate of an electric vehicle according to the present invention;
FIG. 2 is a flowchart of a deep learning-based method for detecting a rider's helmet and a license plate of an electric vehicle according to the present invention;
FIG. 3 is a flow chart of the electric vehicle license plate recognition method based on deep learning provided by the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments 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.
A deep learning-based electric vehicle rider helmet detection and license plate recognition method is shown in figure 1 and comprises the following steps:
step one, self-building an electric vehicle rider helmet and a license plate detection data set; the method comprises the steps of selecting a traffic environment of a traffic intersection as a research background, and selecting scenes of different clothes of riders to shoot and collect pictures of a riding electric vehicle based on spring, summer, autumn and winter. Aiming at the problems that the shot data set is insufficient and the sample is unbalanced, the shot data set needs to be expanded through data enhancement algorithms such as increasing the picture brightness, randomly rotating, reducing the picture brightness, randomly cutting, adding noise and the like. The two data sets jointly form an electric vehicle rider helmet and a license plate detection data set. The data set determines 2-3 detection targets such as helmets, license plates and the like, and open source software labelImg is used as a training set sample labeling tool.
Step two, constructing a detection model based on improved YOLOv5m, and training the model by using the electric vehicle rider helmet and the license plate detection data set; because the helmet and the license plate detected by the model are smaller than the whole picture, a CBAM attention module is added to the Neck part of the YOLOv5m model for detecting the small target and the medium target, so that the YOLOv5m model pays more attention to the channel characteristics of the information of the small target, and the detection capability of the YOLOv5m model for detecting the small target is enhanced. Replacing the GIoU Loss with the CIoU Loss to enable the CIoU Loss to serve as a regression Loss function of the target boundary box, so that the regression stability of the target boundary box is enhanced, and the accuracy of target prediction is improved; replacing the cascading NMS with a DIoU-NMS can minimize the missed detection rate in case of congestion of the target. The method has a good improvement effect on the missing detection of the helmets and license plates with a plurality of electric vehicle scenes in the pictures. And the electric vehicle rider helmet and the license plate detection data set are used for training the model.
Thirdly, segmenting and correcting the electric vehicle license plate in the detection result, and performing data enhancement on the segmented and corrected license plate picture to establish a license plate identification data set; and sending the detection picture into an improved YOLOv5m detection model to obtain a detection result, and segmenting the detected electric vehicle license plate to obtain a complete electric vehicle license plate. However, the license plate of the electric vehicle is inclined, and the complete license plate of the electric vehicle is corrected through affine transformation. The number of the electric vehicle license plates obtained by the method is limited, and a shot data set needs to be expanded by data enhancement algorithms such as increasing picture brightness, randomly rotating, reducing picture brightness, randomly cutting, adding noise and the like. The two data sets jointly form a license plate recognition data set.
Constructing an electric vehicle license plate recognition model based on the improved CRNN, and training the model by using a license plate recognition data set; and fusing the CTPN character detection model and the CRNN character recognition model together, and performing CTPN character detection before performing CRNN character recognition. The CTPN character detection model respectively detects two lines or one line of characters of the electric vehicle license plate and sends the characters into the CRNN electric vehicle license plate recognition model. And constructing an electric vehicle license plate recognition model based on the improved CRNN, and training the model by using a license plate recognition data set.
Inputting the real-time traffic intersection video into the trained model, detecting the condition that the rider does not wear the helmet in the video, and displaying the license plate number of the electric vehicle without the helmet; the method comprises the steps of sending a video stream of a real-time traffic intersection into a detection model based on improved YOLOv5m, detecting the wearing condition of a helmet of an electric vehicle rider in a video, marking the head of the rider and the license plate of the electric vehicle in the video if the rider does not wear the helmet and the electric vehicle has the license plate, sending a detection result based on the improved YOLOv5m detection model into an electric vehicle license plate recognition model based on improved CRNN, recognizing an input license plate, and displaying the recognized license plate number in the video.
As shown in fig. 2, when detecting the helmet and the license plate of the rider of the electric vehicle, the main steps include:
firstly, establishing an electric vehicle rider helmet and a license plate detection data set; the method comprises the steps of selecting a traffic environment of a traffic intersection as a research background, and selecting scenes of different clothes of riders to shoot and collect pictures of a riding electric vehicle based on spring, summer, autumn and winter. Aiming at the problems that the shot data set is insufficient and the sample is unbalanced, the shot data set needs to be expanded through data enhancement algorithms such as increasing the picture brightness, randomly rotating, reducing the picture brightness, randomly cutting, adding noise and the like. The two data sets jointly form an electric vehicle rider helmet and a license plate detection data set.
Step two, constructing a detection model based on improved YOLOv5 m; a CBAM attention module is added at the Neck part of the small target and the medium target detected by the YOLOv5m model, so that the channel characteristics of the small target information in the YOLOv5m model are paid more attention to, and the detection capability of the YOLOv5m model on the small target is enhanced; meanwhile, replacing the GIoU Loss with the CIoU Loss to enable the CIoU Loss to serve as a regression Loss function of the target boundary box, so that the regression stability of the target boundary box is enhanced, and the accuracy of target prediction is improved; replacing the cascading NMS with a DIoU-NMS can minimize the missed detection rate in case of congestion of the target. The method has a good improvement effect on the missing detection of the helmets and license plates with a plurality of electric vehicle scenes in the pictures.
Step three, sending the data set into a model for training; and determining 2-3 detection targets such as helmets, license plates and the like by using the data set, and adopting open source software labelImg as a data set sample labeling tool. The marked self-built electric vehicle rider helmet and license plate detection data set are as follows: 3 into a training set and a test set, and training the model with a pytorreh framework. The YOLOv5m model and the detection model based on the improved YOLOv5m are trained separately with the data sets. The improved YOLOv5 m-based detection model with higher recognition rate is obtained.
Step four, inputting the real-time traffic intersection video into the model; the method comprises the steps of inputting a real-time traffic intersection video based on an improved YOLOv5m detection model, reading the video by the model, cutting the dynamic video into a frame of pictures, and detecting the input pictures by the model in sequence.
Step five, detecting the part of the rider without the helmet and the license plate of the electric vehicle in the video; detecting whether the helmet wearing condition of the rider and the electric vehicle have license plates or not in the video, if the rider does not wear the helmet and the electric vehicle has the license plates, marking the head of the rider and the license plates of the electric vehicle, and displaying the detection result in the video.
As shown in fig. 3, when recognizing a license plate of an electric vehicle, the method mainly includes the following steps:
firstly, establishing a license plate identification data set of an electric vehicle; and sending the detection picture into an improved YOLOv5m detection model to obtain a detection result, and segmenting the detected license plate to obtain a complete license plate of the electric vehicle. However, the license plate of the electric vehicle is inclined, and the complete license plate of the electric vehicle is corrected through affine transformation. The number of the electric vehicle license plates obtained by the method is limited, and a shot data set needs to be expanded by data enhancement algorithms such as increasing picture brightness, randomly rotating, reducing picture brightness, randomly cutting, adding noise and the like. The two data sets jointly form a license plate recognition data set.
Secondly, constructing a license plate recognition model based on the improved CRNN; and fusing the CTPN character detection model and the CRNN character recognition model together, and performing CTPN character detection before performing CRNN character recognition. And constructing an electric vehicle license plate recognition model based on the improved CRNN. The CTPN character detection model respectively detects two lines or one line of characters of the electric vehicle license plate and sends the characters into the CRNN electric vehicle license plate recognition model.
Step three, sending the data set into a model for training; and the data set adopts open source software labelImg as a data set sample labeling tool. The labeled self-built electric vehicle identification data set is as follows 7: and 3, dividing the model into a training set and a testing set in proportion, training the improved CRNN-based license plate recognition model, and obtaining the trained model to prepare for license plate recognition.
Step four, inputting the real-time traffic intersection video into the model; the method comprises the steps of inputting a real-time traffic intersection video based on an improved CRNN license plate recognition model, reading the video by the model, cutting the dynamic video into a frame of pictures, and detecting the input pictures by the model in sequence.
Step five, detecting the part of the rider without the helmet and the license plate of the electric vehicle in the video; and detecting whether the electric vehicle has a license plate or not under the condition that the rider does not wear the helmet in the video. If the electric vehicle has the license plate, the head of the rider and the license plate of the electric vehicle are marked, and the license plate number of the electric vehicle is identified to be displayed in the video.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are given by way of illustration of the principles of the present invention, but that various changes and modifications may be made without departing from the spirit and scope of the invention, and such changes and modifications are within the scope of the invention as claimed.
Claims (8)
1. A deep learning-based electric vehicle rider helmet detection and license plate recognition method is characterized by comprising the following steps:
step one, self-building an electric vehicle rider helmet and a license plate detection data set;
step two, constructing a detection model based on improved YOLOv5m, and training the model by using an electric vehicle rider helmet and a license plate detection data set;
thirdly, segmenting and correcting the electric vehicle license plate in the detection result, and performing data enhancement on the segmented and corrected license plate picture to establish a license plate identification data set;
constructing an electric vehicle license plate recognition model based on the improved CRNN, and training the model by using a license plate recognition data set;
and fifthly, inputting the real-time traffic intersection video into the trained model, detecting the condition that the rider does not wear the helmet in the video, and displaying the license plate number of the electric vehicle without the helmet.
2. The deep learning-based electric vehicle rider helmet detection and license plate recognition method as claimed in claim 1, wherein the collection of the data set in the first step selects a traffic environment at a traffic intersection as a research background, and the pictures of the riding electric vehicle are captured and collected by selecting scenes of different clothes of a rider based on spring, summer, autumn and winter.
3. The method according to claim 1, wherein for the problems of insufficient photographed data set and unbalanced sample in the first step, the photographed data set is expanded by increasing picture brightness, performing random rotation, reducing picture brightness, performing random clipping, adding noise and other data enhancement algorithms to obtain a self-built electric vehicle rider helmet and a license plate detection data set.
4. The deep learning-based electric vehicle rider helmet detection and license plate recognition method as claimed in claim 1, wherein in the second step, a CBAM attention module is added to a Neck part of a YOLOv5m model for detecting a small target and a medium target, so that the YOLOv5m model pays attention to channel characteristics of information of the small target, thereby enhancing the detection capability of the YOLOv5m model for detecting the small target.
5. The deep learning-based electric vehicle rider helmet detection and license plate recognition method according to claim 1, wherein in the second step, the GIoU Loss is replaced by the CIoU Loss, so that the CIoU Loss is used as a regression Loss function of the target boundary frame, thereby enhancing the regression stability of the target boundary frame, improving the accuracy of target prediction, and reducing the missing rate under the condition of target congestion by replacing the cascaded NMS with the DIoU-NMS.
6. The method for detecting the helmet of the rider and identifying the license plate of the electric vehicle based on the deep learning as claimed in claim 1, wherein the image of the license plate is detected in the third step through a modified YOLOv5m detection model, and then the tilted license plate of the electric vehicle is corrected through affine transformation, and the number of the license plate of the electric vehicle needs to be expanded through data enhancement algorithms such as increasing the brightness of the image, randomly rotating the image, reducing the brightness of the image, randomly cutting the image, adding noise and the like.
7. The deep learning-based electric vehicle rider helmet detection and license plate recognition method as claimed in claim 1, wherein the step four of constructing the improved CRNN-based electric vehicle license plate recognition model is to merge a CTPN character detection model with the CRNN character recognition model, perform CTPN character detection before performing CRNN character recognition, and the CTPN character detection model detects two lines or one line of characters of the electric vehicle license plate respectively and sends the two lines or one line of characters into the CRNN electric vehicle license plate recognition model.
8. The method as claimed in claim 1, wherein in the fifth step, the wearing condition of the helmet of the electric vehicle rider in the video is detected based on an improved YOLOv5m detection model, if the rider does not wear the helmet and the electric vehicle has a license plate, the head of the rider and the license plate of the electric vehicle in the video are marked, the detection result based on the improved YOLOv5m detection model is sent to an improved CRNN-based electric vehicle license plate recognition model, the input license plate is recognized, and the recognized license plate is displayed in the video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210750226.8A CN115147869A (en) | 2022-06-28 | 2022-06-28 | Electric vehicle rider helmet detection and license plate recognition method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210750226.8A CN115147869A (en) | 2022-06-28 | 2022-06-28 | Electric vehicle rider helmet detection and license plate recognition method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115147869A true CN115147869A (en) | 2022-10-04 |
Family
ID=83410472
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210750226.8A Pending CN115147869A (en) | 2022-06-28 | 2022-06-28 | Electric vehicle rider helmet detection and license plate recognition method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115147869A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116092017A (en) * | 2023-04-06 | 2023-05-09 | 南京信息工程大学 | Lightweight edge-end vehicle bottom dangerous object identification method, medium and equipment |
-
2022
- 2022-06-28 CN CN202210750226.8A patent/CN115147869A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116092017A (en) * | 2023-04-06 | 2023-05-09 | 南京信息工程大学 | Lightweight edge-end vehicle bottom dangerous object identification method, medium and equipment |
| CN116092017B (en) * | 2023-04-06 | 2023-07-28 | 南京信息工程大学 | Lightweight edge-end vehicle bottom dangerous object identification method, medium and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829403B (en) | Vehicle anti-collision early warning method and system based on deep learning | |
| WO2019196130A1 (en) | Classifier training method and device for vehicle-mounted thermal imaging pedestrian detection | |
| CN109460699B (en) | Driver safety belt wearing identification method based on deep learning | |
| CN105702048B (en) | Highway front truck illegal road occupation identifying system based on automobile data recorder and method | |
| CN104200668B (en) | A kind of motorcycle based on graphical analysis is not helmeted violation event detection method | |
| CN109948418A (en) | A kind of illegal automatic auditing method of violation guiding based on deep learning | |
| CN109949579A (en) | A kind of illegal automatic auditing method that makes a dash across the red light based on deep learning | |
| CN110222596B (en) | Driver behavior analysis anti-cheating method based on vision | |
| Rohith et al. | An efficient helmet detection for MVD using deep learning | |
| CN105760847A (en) | Visual detection method for detecting whether motor cyclist wears helmet or not | |
| CN113160575A (en) | Traffic violation detection method and system for non-motor vehicles and drivers | |
| CN104978567A (en) | Vehicle detection method based on scenario classification | |
| Razalli et al. | Emergency vehicle recognition and classification method using HSV color segmentation | |
| CN109784214A (en) | A kind of detection device and method of railroad track foreign matter | |
| CN114724122B (en) | Target tracking method and device, electronic equipment and storage medium | |
| CN111540171B (en) | Fatigue driving early warning system, corresponding early warning method and construction method | |
| CN115147869A (en) | Electric vehicle rider helmet detection and license plate recognition method based on deep learning | |
| CN112528759A (en) | Traffic violation behavior detection method based on computer vision | |
| CN114926824A (en) | Method for judging bad driving behavior | |
| CN113076852A (en) | Vehicle-mounted snapshot processing system occupying bus lane based on 5G communication | |
| CN112907597A (en) | Railway track line detection method based on deep convolutional neural network | |
| CN109800693B (en) | Night vehicle detection method based on color channel mixing characteristics | |
| CN110782485A (en) | Vehicle lane change detection method and device | |
| CN112633163B (en) | Detection method for realizing illegal operation vehicle detection based on machine learning algorithm | |
| CN202854837U (en) | Vehicle lane mark line detection device |
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 |