CN109948418A - A kind of illegal automatic auditing method of violation guiding based on deep learning - Google Patents
A kind of illegal automatic auditing method of violation guiding based on deep learning Download PDFInfo
- Publication number
- CN109948418A CN109948418A CN201811654644.7A CN201811654644A CN109948418A CN 109948418 A CN109948418 A CN 109948418A CN 201811654644 A CN201811654644 A CN 201811654644A CN 109948418 A CN109948418 A CN 109948418A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- line
- classification
- target vehicle
- illegal
- 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
Abstract
The invention discloses a kind of, and the violation based on deep learning is oriented to illegal automatic auditing method, comprising the following steps: obtains the shorter three different images data of traffic intersection illegal vehicle time interval;All vehicles in three images are detected using detection model;Target vehicle is oriented in the position where three figures using identification model;The lane line and leading line at the crossing where illegal vehicle are partitioned into using parted pattern;Target vehicle driving direction classification is judged using disaggregated model;The driving direction that target vehicle is judged according to target vehicle position and vehicle heading classification results in three figures judges whether vehicle driving method is consistent to judge whether vehicle violates guiding with the segmentation type of the leading line where target vehicle in first figure.The multitask that the present invention realizes detection, segmentation and classification is audited automatically, the manual examination and verification of existing mass data are substituted in this way, manpower is not only greatly saved, accelerates audit speed, while reducing manual examination and verification to a certain extent since tired bring is judged by accident.
Description
Technical field
The present invention relates to the detections of automotive vehicle artificial intelligence, identification, sorting technique field, in particular to a kind of based on deep
The violation of degree study is oriented to illegal automatic auditing system.
Background technique
Constantly improve with living standards of the people with the continuous social and economic development, Urban vehicles poputation rapidly increases
It is long.The illegal examination amount of automotive vehicle also increases rapidly therewith.Traditional vehicle illegal audit is mainly by manually examining
The data of core, magnanimity compare labor intensive by manual examination and verification, and prolonged manpower audit is due to fatigue also influence audit
Efficiency.
How illegal audit accurately and rapidly to be carried out to vehicle, while avoiding manual examination and verification at high cost, fatiguability is easily dredged
The drawbacks such as suddenly, are technical problems urgently to be solved.
Summary of the invention
The purpose of the present invention is: propose that a kind of violation based on deep learning is oriented to illegal automatic auditing method, it is automatic right
Vehicle carries out illegal audit, to meet the needs of nowadays carrying out illegal examination efficiency, accuracy rate to vehicle.
The technical solution adopted by the present invention to solve the technical problems is:
1. a kind of violation based on deep learning is oriented to illegal automatic auditing method, comprising the following steps:
S1, traffic intersection illegal vehicle time interval shorter (1 second to 20 seconds) three different images data are obtained, utilizes road
Mouthful camera capture three of different moments figures from vehicle back because camera is fixed, but target vehicle
Usually in (straight trip is turned left, right-hand rotation) of movement, so first figure obtained is usually to capture from the dead astern of target vehicle
, what second third was captured is likely to be just rear angle or flank angle;
S2, all vehicles in three images are detected using conventional sensing algorithm;
S3, correct vehicle location in first figure is positioned by licence plate recognition method;
S4, vehicle location in above-mentioned S3 is tracked in three figures using GoogLenet network structure;
S5, the crossing being partitioned into using parted pattern where illegal vehicle lane line and leading line;
S6, the classification that target vehicle driving direction is judged using disaggregated model are taken out the target vehicle in three figures,
It is attached to corresponding position in the segmentation result figure of first figure, the training data as train classification models is trained, and data are such as
Shown in Fig. 6;
S7, according to target vehicle position, the position of lane line and vehicle heading classification results in three figures
Judge the driving direction of target vehicle;
S8, judge whether vehicle heading is consistent with the segmentation type of the leading line where target vehicle in first figure
To judge whether vehicle violates guiding.
Using GoogLenet network, to vehicle tracking, steps are as follows:
S40, in training characteristics extraction module, in network, the last one 256 full articulamentum of dimension connects a classification layer, the layer
Classify to different money vehicles, each classification possess the different frame moment acquisition same vehicle, and to the vehicle of all acquisitions into
The enhancing of row data.When trained penalty values loss is reduced to minimum, classification layer is cropped, takes out the upper full connection of one 256 dimension
Layer, 256 dimensional features obtained at this time can be good at characterizing the feature of the vehicle;
S41, GoogLenet Inception-V2 network is input to the vehicle that first figure navigates to, in the network
Input layer carries out padding to the vehicle of input, becomes the consistent image of length and width, extra part is with 0 pixel filling;Then
Up-sampling or down-sampling operation are carried out to pretreated image, unified resize is finally obtained at the image of 200*200 resolution ratio
To 256 dimensional features;
S42, GoogLenet Inception-V2 network is inputted to all vehicles to be matched of second figure, same S41 is obtained
To several 256 dimensional features;
S43, GoogLenet Inception-V2 network is inputted to all vehicles to be matched of third figure, same S41 is obtained
To several 256 dimensional features;
S44, cosine similarity is done with several 256 dimensional features in 256 dimensional features in S41 and S42, since feature mentions
256 dimensional features that modulus block extracts have been able to characterize the vehicle well, so can more show two using cosine is similar
Diversity factor between vehicle finally takes out 256 dimensional features corresponding to highest scoring;
S45, cosine similarity is done with several 256 dimensional features in 256 dimensional features of highest scoring in S42 and S43, taken out
256 dimensional features corresponding to highest scoring;
S46, several vehicles by detection algorithm have been detected due to second figure and third figure respectively, with above-mentioned calculation
Method finds the highest vehicle of similarity score, and taking out vehicle call number corresponding to highest scoring is the vehicle traced into.
The scene cut model obtaining step based on deep learning is as follows:
S51, vehicle in practical application scene, stop line, leading line, the picture of yellow line are collected, and manually marks out these
Region, i.e., artificial mark surround vehicle, stop line, leading line, the closed polygon of yellow line;
S52, artificial mark is converted into label matrix, i.e., all pixels point label in the vehicle closure region manually marked
It is set as 0, all pixels point label is set as 1 in stop line enclosed region, is oriented to all pixels point label in enclosed region and sets
2 are set to, all pixels point label is set as 3 in yellow line enclosed region.
S53, vehicle, stop line, leading line, yellow line picture and the input deeplab-v2 segmentation of corresponding label matrix are calculated
Method training, deeplab-v2 partitioning algorithm is using ResNet-34 as backbone network, psp_module and unet module is as solution
Code device, and skip layer is used to introduce low-dimensional minutia as prototype network structure.Use a*bce_loss+b*
Lovasz_loss introduces auxiliary loss aux_loss and is trained as final loss (0≤a, b≤1 are manually set);
The good deeplab-v2 partitioning algorithm of S54, application training predicts input image pixels point classification, will belong to vehicle,
Stop line, leading line, the pixel coordinate set output of each classification of yellow line, thus realize vehicle, and stop line, leading line, Huang
The segmentation in line region.
The vehicle heading disaggregated model obtaining step based on deep learning is as follows:
S61, take each group traffic intersection capture figure violating the regulations in first be split, obtain segmentation result figure,
Only retain the lane line that segmentation obtains, leading line, stop line, zebra stripes information;
Target vehicle in S62, each group of taking-up figure violating the regulations, and record its location information in three figures;
S63, the corresponding position of segmentation result figure (following Fig. 6 is attached to using the location information of the target vehicle of record
It is shown), and provide the label that target vehicle is straight trip, turns left or turn right;
S64, the disaggregated model training that vehicle heading is carried out using the training dataset of production and its label.
Steps are as follows for the comprehensive descision of the vehicle heading based on lane line, vehicle location and disaggregated model:
S71, the driving direction and score that target vehicle in third figure is provided using vehicle classification model;
The position for the lane line that S72, parted pattern provide, vehicle detection, vehicle identification location model provide in three figures
The position of target vehicle judges the driving direction of vehicle according to the position of target vehicle in the position of lane line and three figures;
S73, the inclined of mutual vehicle relative image y-axis is found out according to the center of target vehicle in three figures respectively
Gyration judges vehicle heading according to deflection angle;
If the result that S74, three kinds of judgment method two of thems provide is consistent, this result as final result, if
The result that three kinds of methods provide is entirely different, and the result score that disaggregated model provides is relatively high, then final result, which uses, divides
It is that class model provides as a result, using the result that is provided using lane line if the result score that disaggregated model provides is relatively low.
The beneficial effects of the present invention are: violating present invention is mainly applied to automotive vehicle, guiding is illegal to be examined automatically, is both saved
About manpower, in turn ensures accuracy.
Detailed description of the invention
Fig. 1 is flow chart of the present invention.
Fig. 2 is structural schematic diagram of the invention.
Fig. 3 is the structural schematic diagram of object detection unit of the present invention.
Fig. 4 is the structural schematic diagram of scene cut unit of the present invention.
Fig. 5 is the structural schematic diagram of vehicle recognition unit of the present invention.
Fig. 6 is that sorter network turns right, keeps straight on, left-hand rotation training data.
Specific embodiment
Below in conjunction with attached drawing, the present invention will be further described.
Present invention is primarily based on module of target detection, vehicle identification locating module, scene cut modules, vehicle heading
Judgment module.
Illegal automatic auditing system is oriented to by object detection unit, vehicle identification positioning unit, field as shown in Fig. 2, violating
Scape cutting unit, vehicle heading judging unit composition.Firstly, image is passed to object detection unit and scene cut list
Member obtains all vehicle locations using target detection model, and it is in place to orient target vehicle institute using vehicle identification location model
It sets.Then, target vehicle is obtained at first using scene cut model according to the approximate location of target vehicle in first figure
Leading line classification where in figure, while being partitioned into all lane line positions.Finally, first recycling vehicle heading classification mould
Type provides the driving direction classification and score of target vehicle in third figure, and combines lane line and target vehicle at three
Position in figure provides the final result of the driving direction of vehicle, is sentenced according to vehicle heading and the leading line classification of segmentation
Whether disconnected vehicle is illegal.
Wherein, not only facilitated without artificial design feature using the method for vehicle heading as described herein judgement but also
Classification accuracy is high, and vehicle heading disaggregated model acquisition methods are as follows:
S1, take each group traffic intersection capture figure violating the regulations in first be split, obtain segmentation result figure, only
Retain the lane line that segmentation obtains, leading line, stop line, zebra stripes information;
Target vehicle in S2, each group of taking-up figure violating the regulations, and record its location information in three figures;
S3, the corresponding position of segmentation result figure (following Fig. 6 institute is attached to using the location information of the target vehicle of record
Show), and provide the label (i.e. 0,1,2) that target vehicle is straight trip, turns left or turn right;
S4, using S1~S3 image generated and corresponding label, (0 to represent target sample be straight trip, and 1 to represent be left
Turn, 2 representatives are to turn right) it partners to form a sample, by the great amount of samples of generation by the CAFFE frame of open source, utilize
Googlenet network training disaggregated model.One passes through trained mould using the image for not knowing label that S1~S3 is generated
Type is exported the result is that an one-dimension array, passes through resulting one-dimension array { score1, score2, score3 }, Ke Yizhi
The driving direction of target vehicle in road image, to achieve the purpose that obtain target vehicle driving direction by disaggregated model.Its
What middle score1 was represented is the score value that target vehicle is straight trip, and size is between 0~1, and what score2 was represented is target vehicle
It is the score value turned left, size is between 0~1, and what score3 was represented is that target vehicle is the score value turned right, and size is 0~1
Between, three values add up to 1, and final result takes maximum corresponding classification in three values, if score3 is maximum in three,
Then final judging result are as follows: target vehicle driving direction is to turn right.
The advantages of basic principles and main features and this programme of this programme have been shown and described above.The technology of the industry
Personnel are it should be appreciated that this programme is not restricted to the described embodiments, and the above embodiments and description only describe this
The principle of scheme, under the premise of not departing from this programme spirit and scope, this programme be will also have various changes and improvements, these changes
Change and improvement is both fallen within the scope of claimed this programme.This programme be claimed range by appended claims and its
Equivalent thereof.
Claims (5)
1. a kind of violation based on deep learning is oriented to illegal automatic auditing method, comprising the following steps:
S1, three different images data that traffic intersection illegal vehicle time interval is 1 second to 20 seconds are obtained, utilizes taking the photograph for crossing
Three for capturing different moments from vehicle back as head scheme;
S2, all vehicles in three images are detected using detection algorithm;
S3, correct vehicle location in first figure is positioned by licence plate recognition method;
S4, vehicle location in above-mentioned S3 is tracked in three figures using GoogLenet network structure;
S5, the crossing being partitioned into using parted pattern where illegal vehicle lane line and leading line;
S6, the driving direction that target vehicle is judged using vehicle heading disaggregated model, by the target vehicle in three figures
It takes out, is attached to corresponding position in the segmentation result figure of first figure, the training data as train classification models is trained;
S7, judged according to target vehicle position, the position of lane line and vehicle heading classification results in three figures
The driving direction of target vehicle;
S8, judge the segmentation type of leading line in vehicle heading and first figure where target vehicle it is whether consistent to
Judge whether vehicle violates guiding.
2. a kind of violation based on deep learning as described in claim 1 is oriented to illegal automatic auditing method, which is characterized in that
It is described that using GoogLenet network, to vehicle tracking, steps are as follows:
S40, in training characteristics extraction module, in network, the last one 256 full articulamentum of dimension connects a classification layer, and the layer is not to
Classify with money vehicle, each classification possesses the same vehicle of different frame moment acquisition, and counts to the vehicle of all acquisitions
According to enhancing, when trained penalty values loss is reduced to minimum, classification layer is cropped, takes out the full articulamentum of upper one 256 dimension, this
When 256 dimensional features that obtain can be good at characterizing the feature of the vehicle;
S41, GoogLenet Inception-V2 network is input to the vehicle that first figure navigates to, in the input of the network
Layer carries out padding to the vehicle of input, becomes the consistent image of length and width, extra part is with 0 pixel filling;Then to pre-
Treated, and image carries out up-sampling or down-sampling operation, and unified resize is finally obtained at the image of 200*200 resolution ratio
One 256 dimensional feature;
S42, GoogLenet Inception-V2 network, same to S41, if obtaining are inputted to all vehicles to be matched of second figure
Dry 256 dimensional features;
S43, GoogLenet Inception-V2 network, same to S41, if obtaining are inputted to all vehicles to be matched of third figure
Dry 256 dimensional features;
S44, cosine similarity is done with several 256 dimensional features in 256 dimensional features in S41 and S42, due to feature extraction mould
256 dimensional features that block extracts have been able to characterize the vehicle well, so can more show two cars using cosine is similar
Between diversity factor, finally take out highest scoring corresponding to 256 dimensional features;
S45, cosine similarity is made of several 256 dimensional features in 256 dimensional features of highest scoring in S42 and S43, take out score
256 dimensional features corresponding to highest;
S46, since second figure and third figure by detection algorithm have detected several vehicles respectively, looked for above-mentioned algorithm
To the highest vehicle of similarity score, taking out vehicle call number corresponding to highest scoring is the vehicle traced into.
3. a kind of violation based on deep learning as described in claim 1 is oriented to illegal automatic auditing method, which is characterized in that
Scene cut model obtaining step of the S5 based on deep learning is as follows:
S51, vehicle in practical application scene, stop line, leading line, the picture of yellow line are collected, and manually mark out these regions,
I.e. artificial mark surrounds vehicle, stop line, leading line, the closed polygon of yellow line;
S52, artificial mark is converted into label matrix, i.e., all pixels point label is arranged in the vehicle closure region manually marked
It is 0, all pixels point label is set as 1 in stop line enclosed region, and all pixels point label is arranged in leading line enclosed region
It is 2, all pixels point label is set as 3 in yellow line enclosed region;
S53, vehicle, stop line, leading line, yellow line picture and corresponding label matrix input deeplab-v2 partitioning algorithm are instructed
Practice, deeplab-v2 partitioning algorithm is using ResNet-34 as backbone network, psp_module and unet module is as decoding
Device, and skip layer is used to introduce low-dimensional minutia as prototype network structure, use a*bce_loss+b*
Lovasz_loss is as final loss, wherein 0≤a, b≤1, and introduce auxiliary loss aux_loss and be trained;
The good deeplab-v2 partitioning algorithm of S54, application training predicts input image pixels point classification, will belong to vehicle, stops
Line, leading line, the pixel coordinate set output of each classification of yellow line, thus realize vehicle, and stop line, leading line, yellow line area
The segmentation in domain.
4. a kind of violation based on deep learning as described in claim 1 is oriented to illegal automatic auditing method, which is characterized in that
Vehicle heading disaggregated model obtaining step of the S6 based on deep learning is as follows:
S61, take each group traffic intersection capture figure violating the regulations in first be split, obtain segmentation result figure, only protect
Stay the lane line that segmentation obtains, leading line, stop line, zebra stripes information;
Target vehicle in S62, each group of taking-up figure violating the regulations, and record its location information in three figures;
S63, it is attached to the corresponding position of segmentation result figure using the location information of the target vehicle of record, and provides target carriage
It is straight trip, the label that turns left or turn right;
S64, the disaggregated model training that vehicle heading is carried out using the training dataset of production and its label.
5. a kind of violation based on deep learning as described in claim 1 is oriented to illegal automatic auditing method, which is characterized in that
Steps are as follows for the comprehensive descision of the vehicle heading based on lane line, vehicle location and disaggregated model:
S71, the driving direction and score that target vehicle in third figure is provided using vehicle classification model;
The position for the lane line that S72, parted pattern provide, vehicle detection, vehicle identification location model provide target in three figures
The position of vehicle judges the driving direction of vehicle according to the position of target vehicle in the position of lane line and three figures;
S73, the deflection angle for finding out mutual vehicle relative image y-axis respectively according to the center of target vehicle in three figures
Degree, judges vehicle heading according to deflection angle;
If the result that S74, three kinds of judgment method two of thems provide is consistent, this result as final result, if three kinds
The result that method provides is entirely different, and the result score that disaggregated model provides is relatively high, then final result is using classification mould
It is that type provides as a result, using the result that is provided using lane line if the result score that disaggregated model provides is relatively low.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811654644.7A CN109948418A (en) | 2018-12-31 | 2018-12-31 | A kind of illegal automatic auditing method of violation guiding based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811654644.7A CN109948418A (en) | 2018-12-31 | 2018-12-31 | A kind of illegal automatic auditing method of violation guiding based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109948418A true CN109948418A (en) | 2019-06-28 |
Family
ID=67007213
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811654644.7A Pending CN109948418A (en) | 2018-12-31 | 2018-12-31 | A kind of illegal automatic auditing method of violation guiding based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109948418A (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110415529A (en) * | 2019-09-04 | 2019-11-05 | 上海眼控科技股份有限公司 | Automatic processing method, device, computer equipment and the storage medium of vehicle violation |
CN110458031A (en) * | 2019-07-15 | 2019-11-15 | 邱玉枝 | The recognition methods of vehicle violation and device |
CN110717433A (en) * | 2019-09-30 | 2020-01-21 | 华中科技大学 | Deep learning-based traffic violation analysis method and device |
CN110909699A (en) * | 2019-11-28 | 2020-03-24 | 北京以萨技术股份有限公司 | Video vehicle non-guide driving detection method and device and readable storage medium |
CN111091041A (en) * | 2019-10-22 | 2020-05-01 | 上海眼控科技股份有限公司 | Vehicle law violation judging method and device, computer equipment and storage medium |
CN111126213A (en) * | 2019-12-13 | 2020-05-08 | 苏州智加科技有限公司 | Lane line detection method and device based on historical cache data and storage medium |
CN111259767A (en) * | 2020-01-13 | 2020-06-09 | 厦门大学 | Traffic illegal behavior identification method and system based on traffic data and street view data |
CN111339834A (en) * | 2020-02-04 | 2020-06-26 | 浙江大华技术股份有限公司 | Method for recognizing vehicle traveling direction, computer device, and storage medium |
CN111368774A (en) * | 2020-03-12 | 2020-07-03 | 北京以萨技术股份有限公司 | Waste film rollback method, system, terminal and medium based on traffic violation image |
CN111563463A (en) * | 2020-05-11 | 2020-08-21 | 上海眼控科技股份有限公司 | Method and device for identifying road lane lines, electronic equipment and storage medium |
CN112101268A (en) * | 2020-09-23 | 2020-12-18 | 浙江浩腾电子科技股份有限公司 | Vehicle line pressing detection method based on geometric projection |
CN112699827A (en) * | 2021-01-05 | 2021-04-23 | 长威信息科技发展股份有限公司 | Traffic police affair handling method and system based on block chain |
CN113327414A (en) * | 2020-02-28 | 2021-08-31 | 深圳市丰驰顺行信息技术有限公司 | Vehicle reverse running detection method and device, computer equipment and storage medium |
CN114693722A (en) * | 2022-05-31 | 2022-07-01 | 山东极视角科技有限公司 | Vehicle driving behavior detection method, detection device and detection equipment |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015089867A1 (en) * | 2013-12-17 | 2015-06-25 | 东莞中国科学院云计算产业技术创新与育成中心 | Traffic violation detection method |
CN107730904A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Multitask vehicle driving in reverse vision detection system based on depth convolutional neural networks |
-
2018
- 2018-12-31 CN CN201811654644.7A patent/CN109948418A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015089867A1 (en) * | 2013-12-17 | 2015-06-25 | 东莞中国科学院云计算产业技术创新与育成中心 | Traffic violation detection method |
CN107730904A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Multitask vehicle driving in reverse vision detection system based on depth convolutional neural networks |
Non-Patent Citations (1)
Title |
---|
刘海等: "基于OpenCV和机器学习的违法停车检测算法", 《上海船舶运输科学研究所学报》 * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110458031A (en) * | 2019-07-15 | 2019-11-15 | 邱玉枝 | The recognition methods of vehicle violation and device |
CN110415529A (en) * | 2019-09-04 | 2019-11-05 | 上海眼控科技股份有限公司 | Automatic processing method, device, computer equipment and the storage medium of vehicle violation |
CN110415529B (en) * | 2019-09-04 | 2021-09-28 | 上海眼控科技股份有限公司 | Automatic processing method and device for vehicle violation, computer equipment and storage medium |
CN110717433A (en) * | 2019-09-30 | 2020-01-21 | 华中科技大学 | Deep learning-based traffic violation analysis method and device |
CN111091041A (en) * | 2019-10-22 | 2020-05-01 | 上海眼控科技股份有限公司 | Vehicle law violation judging method and device, computer equipment and storage medium |
CN110909699A (en) * | 2019-11-28 | 2020-03-24 | 北京以萨技术股份有限公司 | Video vehicle non-guide driving detection method and device and readable storage medium |
CN111126213A (en) * | 2019-12-13 | 2020-05-08 | 苏州智加科技有限公司 | Lane line detection method and device based on historical cache data and storage medium |
CN111126213B (en) * | 2019-12-13 | 2022-09-02 | 苏州智加科技有限公司 | Lane line detection method and device based on historical cache data and storage medium |
CN111259767A (en) * | 2020-01-13 | 2020-06-09 | 厦门大学 | Traffic illegal behavior identification method and system based on traffic data and street view data |
CN111259767B (en) * | 2020-01-13 | 2023-04-18 | 厦门大学 | Traffic illegal behavior identification method and system based on traffic data and street view data |
CN111339834A (en) * | 2020-02-04 | 2020-06-26 | 浙江大华技术股份有限公司 | Method for recognizing vehicle traveling direction, computer device, and storage medium |
CN113327414A (en) * | 2020-02-28 | 2021-08-31 | 深圳市丰驰顺行信息技术有限公司 | Vehicle reverse running detection method and device, computer equipment and storage medium |
CN111368774A (en) * | 2020-03-12 | 2020-07-03 | 北京以萨技术股份有限公司 | Waste film rollback method, system, terminal and medium based on traffic violation image |
CN111563463A (en) * | 2020-05-11 | 2020-08-21 | 上海眼控科技股份有限公司 | Method and device for identifying road lane lines, electronic equipment and storage medium |
CN112101268A (en) * | 2020-09-23 | 2020-12-18 | 浙江浩腾电子科技股份有限公司 | Vehicle line pressing detection method based on geometric projection |
CN112101268B (en) * | 2020-09-23 | 2022-07-29 | 浙江浩腾电子科技股份有限公司 | Vehicle line pressing detection method based on geometric projection |
CN112699827A (en) * | 2021-01-05 | 2021-04-23 | 长威信息科技发展股份有限公司 | Traffic police affair handling method and system based on block chain |
CN112699827B (en) * | 2021-01-05 | 2023-07-25 | 长威信息科技发展股份有限公司 | Traffic police treatment method and system based on blockchain |
CN114693722A (en) * | 2022-05-31 | 2022-07-01 | 山东极视角科技有限公司 | Vehicle driving behavior detection method, detection device and detection equipment |
CN114693722B (en) * | 2022-05-31 | 2022-09-09 | 山东极视角科技有限公司 | Vehicle driving behavior detection method, detection device and detection equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
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 | |
CN109637151A (en) | A kind of recognition methods that highway Emergency Vehicle Lane is driven against traffic regulations | |
CN104751634B (en) | The integrated application method of freeway tunnel driving image acquisition information | |
CN109948416A (en) | A kind of illegal occupancy bus zone automatic auditing method based on deep learning | |
CN105702048B (en) | Highway front truck illegal road occupation identifying system based on automobile data recorder and method | |
CN105070053B (en) | A kind of intelligent traffic monitoring video camera for recognizing rule-breaking vehicle motor pattern | |
CN102903239B (en) | Method and system for detecting illegal left-and-right steering of vehicle at traffic intersection | |
CN102867417B (en) | Taxi anti-forgery system and taxi anti-forgery method | |
CN108009518A (en) | A kind of stratification traffic mark recognition methods based on quick two points of convolutional neural networks | |
CN110689724B (en) | Automatic motor vehicle zebra crossing present pedestrian auditing method based on deep learning | |
CN107085696A (en) | A kind of vehicle location and type identifier method based on bayonet socket image | |
CN109948417A (en) | A kind of vehicle based on deep learning drives in the wrong direction illegal automatic auditing method | |
CN109033950A (en) | Vehicle based on multiple features fusion cascade deep model, which is disobeyed, stops detection method | |
CN109766769A (en) | A kind of road target detection recognition method based on monocular vision and deep learning | |
CN103824037B (en) | Vehicle anti-tracking alarm device | |
Han et al. | Deep learning-based workers safety helmet wearing detection on construction sites using multi-scale features | |
CN106919939B (en) | A kind of traffic signboard tracks and identifies method and system | |
CN106934374A (en) | The recognition methods of traffic signboard and system in a kind of haze scene | |
CN113160575A (en) | Traffic violation detection method and system for non-motor vehicles and drivers | |
CN110378243A (en) | A kind of pedestrian detection method and device | |
CN110516633A (en) | A kind of method for detecting lane lines and system based on deep learning | |
CN111414861A (en) | Method for realizing detection processing of pedestrians and non-motor vehicles based on deep learning | |
CN113450573A (en) | Traffic monitoring method and traffic monitoring system based on unmanned aerial vehicle image recognition | |
CN115965926B (en) | Vehicle-mounted road sign marking inspection system |
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 | ||
AD01 | Patent right deemed abandoned | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20230721 |