CN109670450A - A kind of people's vehicle object detecting method based on video - Google Patents
A kind of people's vehicle object detecting method based on video Download PDFInfo
- Publication number
- CN109670450A CN109670450A CN201811565548.5A CN201811565548A CN109670450A CN 109670450 A CN109670450 A CN 109670450A CN 201811565548 A CN201811565548 A CN 201811565548A CN 109670450 A CN109670450 A CN 109670450A
- Authority
- CN
- China
- Prior art keywords
- people
- method based
- detecting method
- object detecting
- vehicle object
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- 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
-
- 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
-
- 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
-
- 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
- G06V2201/08—Detecting or categorising vehicles
-
- 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
Abstract
The present invention provides a kind of people's vehicle object detecting method based on video, comprising the following steps: A. data acquisition；B. the data in step A are labeled；C. training set and test set are generated to the data after mark；D. building convolutional neural networks E. carries out model training to the convolutional neural networks of step D；F. it detects.The invention has the advantages that: verification and measurement ratio height, and for some vehicles for being difficult to be identified with car plate detection, candid photograph can reach good effect；Identification monitoring for pedestrian and non motorized vehicle, also there is more accurately recognition effect, can preferably realize various monitoring evidence obtainings, for harmonious society, safety traffic, intelligent travel provides a guarantee.
Description
Technical field
The invention belongs to traffic monitoring technical fields, more particularly, to a kind of people's vehicle object detecting method based on video.
Background technique
In field of traffic, inevitably the detection of vehicle, pedestrian, non-motor vehicle is separated, is then supervised respectively
Control, carries out early warning and the record of illegal incidents, it may be said that be that the core of traffic monitoring technical field compares vehicle detection
Mature is the vehicle detection based on license plate, and accuracy can be up to 99%, but for some unlicensed vehicles and some works
Journey vehicle can not capture vehicle by Car license recognition effective position, cause the difficulty of some post-mordem forensics work in this way；Pedestrian and non-
Motor vehicle detecting, since target is relatively small, and posture feature complexity is much higher than vehicle detection, up to the present, also still
It is old to be explored in optimization in continuous；Pedestrian and non motorized vehicle is effectively detected into equally as the main object in traffic
It is divided into essential part, for the progress important role of field of traffic.
Summary of the invention
In view of this, the present invention is directed to propose a kind of people's vehicle object detecting method based on video, with what is solved the above problems
Shortcoming.
In order to achieve the above objectives, the technical scheme of the present invention is realized as follows:
A kind of people's vehicle object detecting method based on video, comprising the following steps:
A. data acquire；
B. the data in step A are labeled；
C. training set and test set are generated to the data after mark；
D. convolutional neural networks are constructed；
E. model training is carried out to the convolutional neural networks of step D；
F. it detects.
Further, it is acquired in the step A and contains all kinds of traffic targets in the different time sections of various traffic scenes
Picture.
Further, it is labeled using the boundary rectangle of traffic target as boundary in the step B.
Further, random to generate training set and test set according to the ratio of number of pictures 4:1 in the step C.
Further, the process that convolutional neural networks are constructed in the step D is as follows:
D1. VGG network is used, the filter and interval in volume base with multiple sizes less than or equal to 3*3 replace primary
Property filter using size more than or equal to 5*5 and interval；
D2. it is respective below to remove every layer of Conv1_2, Conv2_2, Conv3_2, Conv4_2, Conv5_2 in VGG network
The pond layer of connection, and Conv5_x is added, Conv6_x, Conv7_x, tetra- groups of convolution modules of Conv8_x, wherein each convolution
Convy_1 is the 1/2 of the channel number of Convy_2 in module group.
Further, the training process in the step E is as follows:
E1. by change brightness, saturation degree, rotation, mirror image and by image cropping to the training set generated in step C
Data enhancing is carried out with test set；
E2. above the feature map obtained from Conv4_2, Conv5_2, Conv6_2, Conv7_2, Conv8_2 respectively
It carries out position recurrence and class probability calculates.
Further, the detection process in the step F is as follows:
F1. image color switching: being converted to BGR format by YUV for color of image format, and switch condition is as follows,
B=Y+1.779* (U-128)
G=Y-0.3455* (U-128) -0.7169* (V-128)
R=Y+1.4075* (V-128)；
F2. BGR format is sent into trained model to detect, model output detects obtained target type, mesh
Cursor position (x, y, w, h) and objective degrees of confidence；
F3. result filters, and by limited target confidence level, filters out erroneous detection target；By handing over and than limiting, it is most accurate to obtain
Target type information；By non-maxima suppression, most accurate target position information (x, y, w, h) is obtained.
Compared with the existing technology, people's vehicle object detecting method of the present invention based on video has the advantage that
People's vehicle object detecting method verification and measurement ratio of the present invention based on video is high, is difficult to be known with car plate detection for some
Other vehicle, candid photograph can reach good effect；Identification monitoring for pedestrian and non motorized vehicle, also has and more accurately identifies
Effect can preferably realize various monitoring evidence obtainings, and for harmonious society, safety traffic, intelligent travel provides a guarantee.
Detailed description of the invention
The attached drawing for constituting a part of the invention is used to provide further understanding of the present invention, schematic reality of the invention
It applies example and its explanation is used to explain the present invention, do not constitute improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is training flow chart described in the embodiment of the present invention；
Fig. 2 is the network structure of the embodiment of the present invention.
Specific embodiment
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention can phase
Mutually combination.
The present invention will be described in detail below with reference to the accompanying drawings and embodiments.
As shown in Figure 1, a kind of people's vehicle object detecting method based on video, comprising the following steps:
A. data acquire；
B. the data in step A are labeled；
C. training set and test set are generated to the data after mark；
D. convolutional neural networks are constructed；
E. model training is carried out to the convolutional neural networks of step D；
F. it detects.
The different time sections (including daytime, night, frontlighting, backlight) that various traffic scenes are acquired in the step A include
There is the picture of all kinds of traffic targets, wherein all kinds of traffic targets include Bicycle (non-maneuver cart), Pedestrain (row
People), Car (car), Truck (lorry), Bus (bus), Tricycle (tricycle), Engineering_Truck (work
Journey vehicle)
It is labeled using the boundary rectangle of traffic target as boundary in the step B, label is Bicycle (non-maneuver two
Take turns vehicle), Pedestrain (pedestrian), Car (car), Truck (lorry), Bus (bus), Tricycle (tricycle),
Engineering_Truck (engineering truck).
It is random to generate training set and test set according to the ratio of number of pictures 4:1 in the step C；The effect of training set
It is trained network weight, the effect of test set is to test the accuracy of training pattern, prevents from training over-fitting.
The process that convolutional neural networks are constructed in the step D is as follows:
D1. VGG network is used, replaces a disposable big ruler with several small size filter and interval in volume base
Very little filter and interval, wherein small size generally refers to 3*3 and filter below；Large scale refers generally to 5*5's or more
Size, does not change the visual field of input data in this way, and comprehensive design goes out improved convolutional network for the training of people's vehicle object, two 3*3
Convolution can be consistent with the convolution receptive field of a 5*5, the convolution of three 3*3 can be with the convolution impression of a 7*7 also one
It causes；
D2. it is respective below to remove every layer of Conv1_2, Conv2_2, Conv3_2, Conv4_2, Conv5_2 in VGG network
The pond layer of connection reduces the loss of information caused by dimensionality reduction, and adds Conv5_x, Conv6_x, Conv7_x, Conv8_x tetra-
Group convolution module, wherein Convy_1 is the 1/2 of the channel number of Convy_2 in each convolution module group, by deepening convolution
Network, the different configurations of channel are available to arrive more thoroughgoing and painstaking target signature；Specific network structure such as Fig. 2 institute
Show；
Training process in the step E is as follows:
E1. by changing brightness, saturation degree, rotation, mirror image and passing through image cropping to the (life in step C of existing sample
At training set and test set) carry out data enhancing, since deep learning needs the number of samples of substantial amounts, need to existing
Sample carries out data enhancing processing, by change brightness, saturation degree, rotation, mirror image and passes through the progress of the methods of image cropping
Data enhancing；
E2. above the feature map obtained from Conv4_2, Conv5_2, Conv6_2, Conv7_2, Conv8_2 respectively
Position recurrence and class probability is carried out to calculate, in order to enable model has Analysis On Multi-scale Features, from Conv4_2, Conv5_2,
Position recurrence is carried out above the feature map that Conv6_2, Conv7_2, Conv8_2 are obtained respectively and class probability calculates.
Even in this way, also have the significant position feature of comparison on biggish characteristic pattern compared with Small object, thus can be compatible with and meanwhile compared with
The good big target of detection and Small object,
The loss function Loss used is the weighting of location error mbox_loc and confidence level error mbox_conf:
Loss (x, c, l, g)=1/N (mbox_conf (x, c)+α * mbox_loc (x, l, g))
Wherein, weight proportion α value range is 0-1, and N is Ground Truth (the true frame of target) positive sample quantity, and c is
Classification confidence level predicted value, l are the position prediction value of prediction block, and g is the location parameter of Ground Truth；
Location error mbox_loc is calculated using SmoothL1loss:
Wherein, k indicates Ground Truth generic,Indicate i-th of prediction block and j-th of true frame about class
Whether other k matches, otherwise it is 0 that matching, which is 1,；
Confidence level error mbox_conf is obtained by softmaxloss method:
Wherein, p indicates the classification of prediction,Indicate i-th of prediction block and j-th true frame about classification p whether
Match.
Detection process in the step F is as follows:
F1. image color switching: since the picture format that monitoring camera is got is YUV type, and network inputs are BGR
Format, therefore color of image format is converted into BGR format by YUV, switch condition is as follows,
B=Y+1.779* (U-128)
G=Y-0.3455* (U-128) -0.7169* (V-128)
R=Y+1.4075* (V-128)
F2. BGR format is sent into trained model to detect, model output detects obtained target type, mesh
Cursor position (x, y, w, h) and objective degrees of confidence；
F3. result filters, and by limited target confidence level, filters out erroneous detection target；By handing over and than (Intersection-
Over-Union, i.e. IOU) it limits, obtain most accurate Target type information；Pass through non-maxima suppression (Non-Maximum
Suppression, i.e. NMS), obtain most accurate target position information (x, y, w, h).In the present embodiment, take confidence level big
In or to be equal to 0.8 be correct target, that the one target less than 0.8 is then filtered out；Take IOU be greater than or equal to 0.4, then it is assumed that be for
Same target；Taking NMS threshold value is 0.4.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (7)
1. a kind of people's vehicle object detecting method based on video, which comprises the following steps:
A. data acquire；
B. the data in step A are labeled；
C. training set and test set are generated to the data after mark；
D. convolutional neural networks are constructed；
E. model training is carried out to the convolutional neural networks of step D；
F. it detects.
2. a kind of people's vehicle object detecting method based on video according to claim 1, it is characterised in that: in the step A
Acquire the picture for containing all kinds of traffic targets in the different time sections of various traffic scenes.
3. a kind of people's vehicle object detecting method based on video according to claim 1, it is characterised in that: in the step B
It is labeled using the boundary rectangle of traffic target as boundary.
4. a kind of people's vehicle object detecting method based on video according to claim 1, it is characterised in that: in the step C
It is random to generate training set and test set according to the ratio of number of pictures 4:1.
5. a kind of people's vehicle object detecting method based on video according to claim 1, which is characterized in that in the step D
The process for constructing convolutional neural networks is as follows:
D1. VGG network is used, the filter and interval in volume base with multiple sizes less than or equal to 3*3 replace disposably making
Filter and interval with a size more than or equal to 5*5；
D2. remove every layer of Conv1_2, Conv2_2, Conv3_2, Conv4_2, Conv5_2 respectively connection below in VGG network
Pond layer, and add Conv5_x, Conv6_x, Conv7_x, tetra- groups of convolution modules of Conv8_x, wherein each convolution module
Convy_1 is the 1/2 of the channel number of Convy_2 in group.
6. a kind of people's vehicle object detecting method based on video according to claim 1, which is characterized in that in the step E
Training process it is as follows:
E1. by change brightness, saturation degree, rotation, mirror image and by image cropping to the training set and survey generated in step C
Examination collection carries out data enhancing；
E2. it is carried out respectively above the feature map obtained from Conv4_2, Conv5_2, Conv6_2, Conv7_2, Conv8_2
Position returns and class probability calculates.
7. a kind of people's vehicle object detecting method based on video according to claim 1, which is characterized in that in the step F
Detection process it is as follows:
Color of image format is converted to BGR format by YUV by F1. image color switching, and switch condition is as follows,
B=Y+1.779* (U-128)
G=Y-0.3455* (U-128) -0.7169* (V-128)
R=Y+1.4075* (V-128)；
F2. BGR format is sent into trained model to detect, model output detects obtained target type, target position
Set (x, y, w, h) and objective degrees of confidence；
F3. result filters, and by limited target confidence level, filters out erroneous detection target；By handing over and than limiting, most accurate mesh is obtained
Mark type information；By non-maxima suppression, most accurate target position information (x, y, w, h) is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811565548.5A CN109670450B (en) | 2018-12-20 | 2018-12-20 | Video-based man-vehicle object detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811565548.5A CN109670450B (en) | 2018-12-20 | 2018-12-20 | Video-based man-vehicle object detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109670450A true CN109670450A (en) | 2019-04-23 |
| CN109670450B CN109670450B (en) | 2023-07-25 |
Family
ID=66144134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811565548.5A Active CN109670450B (en) | 2018-12-20 | 2018-12-20 | Video-based man-vehicle object detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109670450B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110111577A (en) * | 2019-05-15 | 2019-08-09 | 武汉纵横智慧城市股份有限公司 | Non-motor vehicle recognition methods, device, equipment and storage medium based on big data |
| CN110399800A (en) * | 2019-06-28 | 2019-11-01 | 智慧眼科技股份有限公司 | Detection method of license plate and system, storage medium based on deep learning VGG16 frame |
| CN111461130A (en) * | 2020-04-10 | 2020-07-28 | 视研智能科技(广州)有限公司 | High-precision image semantic segmentation algorithm model and segmentation method |
| CN113239962A (en) * | 2021-04-12 | 2021-08-10 | 南京速度软件技术有限公司 | Traffic participant identification method based on single fixed camera |
| CN114973154A (en) * | 2022-07-29 | 2022-08-30 | 成都宜泊信息科技有限公司 | Parking lot identification method, parking lot identification system, parking lot control method, parking lot control system, parking lot equipment and parking lot control medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104537387A (en) * | 2014-12-16 | 2015-04-22 | 广州中国科学院先进技术研究所 | Method and system for classifying automobile types based on neural network |
| US20170140253A1 (en) * | 2015-11-12 | 2017-05-18 | Xerox Corporation | Multi-layer fusion in a convolutional neural network for image classification |
| CN107316007A (en) * | 2017-06-07 | 2017-11-03 | 浙江捷尚视觉科技股份有限公司 | A kind of monitoring image multiclass object detection and recognition methods based on deep learning |
| CN107358182A (en) * | 2017-06-29 | 2017-11-17 | 维拓智能科技（深圳）有限公司 | Pedestrian detection method and terminal device |
| CN107944442A (en) * | 2017-11-09 | 2018-04-20 | 北京智芯原动科技有限公司 | Based on the object test equipment and method for improving convolutional neural networks |
| CN108090457A (en) * | 2017-12-27 | 2018-05-29 | 天津天地人和企业管理咨询有限公司 | A kind of motor vehicle based on video does not give precedence to pedestrian detection method |
| CN108564555A (en) * | 2018-05-11 | 2018-09-21 | 中北大学 | A kind of digital image noise reduction method based on NSST and CNN |
-
2018
- 2018-12-20 CN CN201811565548.5A patent/CN109670450B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104537387A (en) * | 2014-12-16 | 2015-04-22 | 广州中国科学院先进技术研究所 | Method and system for classifying automobile types based on neural network |
| US20170140253A1 (en) * | 2015-11-12 | 2017-05-18 | Xerox Corporation | Multi-layer fusion in a convolutional neural network for image classification |
| CN107316007A (en) * | 2017-06-07 | 2017-11-03 | 浙江捷尚视觉科技股份有限公司 | A kind of monitoring image multiclass object detection and recognition methods based on deep learning |
| CN107358182A (en) * | 2017-06-29 | 2017-11-17 | 维拓智能科技（深圳）有限公司 | Pedestrian detection method and terminal device |
| CN107944442A (en) * | 2017-11-09 | 2018-04-20 | 北京智芯原动科技有限公司 | Based on the object test equipment and method for improving convolutional neural networks |
| CN108090457A (en) * | 2017-12-27 | 2018-05-29 | 天津天地人和企业管理咨询有限公司 | A kind of motor vehicle based on video does not give precedence to pedestrian detection method |
| CN108564555A (en) * | 2018-05-11 | 2018-09-21 | 中北大学 | A kind of digital image noise reduction method based on NSST and CNN |
Non-Patent Citations (4)
| Title |
|---|
| 刘明明，张敏情，刘佳等: ""一种基于浅层卷积神经网络的隐写分析方法"", 《山东大学学报（理学版）》 * |
| 刘明明，张敏情，刘佳等: ""一种基于浅层卷积神经网络的隐写分析方法"", 《山东大学学报（理学版）》, vol. 53, no. 03, 20 March 2018 (2018-03-20), pages 63 - 70 * |
| 贾瑞明,刘立强: ""基于卷积神经网络的交通灯识别"", 《北方工业大学学报》 * |
| 贾瑞明,刘立强: ""基于卷积神经网络的交通灯识别"", 《北方工业大学学报》, vol. 30, no. 05, 31 October 2018 (2018-10-31), pages 22 - 30 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110111577A (en) * | 2019-05-15 | 2019-08-09 | 武汉纵横智慧城市股份有限公司 | Non-motor vehicle recognition methods, device, equipment and storage medium based on big data |
| CN110399800A (en) * | 2019-06-28 | 2019-11-01 | 智慧眼科技股份有限公司 | Detection method of license plate and system, storage medium based on deep learning VGG16 frame |
| CN110399800B (en) * | 2019-06-28 | 2021-05-07 | 智慧眼科技股份有限公司 | License plate detection method and system based on deep learning VGG16 framework and storage medium |
| CN111461130A (en) * | 2020-04-10 | 2020-07-28 | 视研智能科技(广州)有限公司 | High-precision image semantic segmentation algorithm model and segmentation method |
| CN111461130B (en) * | 2020-04-10 | 2021-02-09 | 视研智能科技(广州)有限公司 | High-precision image semantic segmentation algorithm model and segmentation method |
| CN113239962A (en) * | 2021-04-12 | 2021-08-10 | 南京速度软件技术有限公司 | Traffic participant identification method based on single fixed camera |
| CN114973154A (en) * | 2022-07-29 | 2022-08-30 | 成都宜泊信息科技有限公司 | Parking lot identification method, parking lot identification system, parking lot control method, parking lot control system, parking lot equipment and parking lot control medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109670450B (en) | 2023-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111368687B (en) | Sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation | |
| CN109670450A (en) | A kind of people's vehicle object detecting method based on video | |
| CN107066953B (en) | A kind of vehicle cab recognition towards monitor video, tracking and antidote and device | |
| US10614327B2 (en) | Traffic-light detection and classification using computer vision and deep learning | |
| CN110969160B (en) | License plate image correction and recognition method and system based on deep learning | |
| CN103279756B (en) | Vehicle detection based on integrated classifier analyzes system and determination method thereof | |
| CN103942560B (en) | A kind of high-resolution video vehicle checking method in intelligent traffic monitoring system | |
| CN109948416A (en) | A kind of illegal occupancy bus zone automatic auditing method based on deep learning | |
| Suhweil et al. | Smart controlling for traffic light time | |
| CN106934374B (en) | Method and system for identifying traffic signboard in haze scene | |
| CN109409242A (en) | A kind of black smoke vehicle detection method based on cyclic convolution neural network | |
| CN103093249A (en) | Taxi identifying method and system based on high-definition video | |
| Yonetsu et al. | Two-stage YOLOv2 for accurate license-plate detection in complex scenes | |
| Tourani et al. | Vehicle counting method based on digital image processing algorithms | |
| CN110069982A (en) | A kind of automatic identifying method of vehicular traffic and pedestrian | |
| Pflugfelder et al. | On learning vehicle detection in satellite video | |
| CN115424217A (en) | AI vision-based intelligent vehicle identification method and device and electronic equipment | |
| CN106557754A (en) | A kind of vehicle detection at night and state judging method | |
| Tao et al. | Smoky vehicle detection based on range filtering on three orthogonal planes and motion orientation histogram | |
| CN115376108A (en) | Obstacle detection method and device in complex weather | |
| Pillai et al. | Vehicle type and color classification and detection for amber and silver alert emergencies using machine learning | |
| Krishnakumar et al. | Detection of vehicle speeding violation using video processing techniques | |
| CN114359196A (en) | Fog detection method and system | |
| CN111724607A (en) | Steering lamp use detection method and device, computer equipment and storage medium | |
| Guo et al. | Visibility detection based on the recognition of the preceding vehicle’s taillight signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |