CN106408932A - Mobile terminal based pre-warning system for distractive walks - Google Patents

Mobile terminal based pre-warning system for distractive walks Download PDF

Info

Publication number
CN106408932A
CN106408932A CN201610881571.XA CN201610881571A CN106408932A CN 106408932 A CN106408932 A CN 106408932A CN 201610881571 A CN201610881571 A CN 201610881571A CN 106408932 A CN106408932 A CN 106408932A
Authority
CN
China
Prior art keywords
user
mobile terminal
sample
unit
decision
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
Application number
CN201610881571.XA
Other languages
Chinese (zh)
Other versions
CN106408932B (en
Inventor
王晓亮
阮锦绣
唐茂志
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University
Original Assignee
Nanjing University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University filed Critical Nanjing University
Priority to CN201610881571.XA priority Critical patent/CN106408932B/en
Publication of CN106408932A publication Critical patent/CN106408932A/en
Application granted granted Critical
Publication of CN106408932B publication Critical patent/CN106408932B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/005Traffic control systems for road vehicles including pedestrian guidance indicator
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/10Terrestrial scenes

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Multimedia (AREA)
  • Traffic Control Systems (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention discloses a mobile terminal based pre-warning system for distractive walks, comprising a model training module and a real time detection module. The model training module conducts pre-treatment and model training based on the large amount of training data it has collected to generate a sample model for real time detection. The real time detection module, through the use of the mobile terminal, acquires the sample and conducts the pre-treatment. Based on the sample model generated by the model training module, the real time detection module conducts separate detections to the acquired samples. According to the results of the separate detections, decisions and judgments are made so that warnings can be made to a distractive user who walks out of a sidewalk. According to the invention, it is possible to make warnings to a user in advance before the user walks towards a dangerous zone. Without extra hardware arrangement, effective warnings can be made to the mobile terminal of a distractive walker, and the previous detection results can be effectively utilized. With high accuracy, the pre-warning system consumes less energy and can make judgments in advance.

Description

A kind of " walking of diverting one's attention " early warning system based on mobile terminal
Technical field
The present invention relates to walking security fields of diverting one's attention, especially a kind of " walking of diverting one's attention " early warning system based on mobile terminal System.
Background technology
In actual life, pedestrian's wholeheartedly dual-purpose, that is, the time-division heart of walking is sent short messages on mobile phone, makes a phone call etc. thus causing peace The phenomenon of full problem is frequent all the more.One research coming from Stony Brook University SUNY is pointed out, diverts one's attention to send out short during walking The people of letter 60% probability more than the people of normal walking meets with vehicle accident.A report of N.S.C. Point out, between 2000 to 2011, pedestrian because mobile phone divert attention in accident estimated cause altogether injured more than 10,000 people.Prior art In, continued to bring out using the technology that the various sensors on mobile terminal solve this problem, such as judge position using GPS, profit Judge mobile phone state with accelerator, gyroscope, judge car etc. using mike, but all there is certain defect, such as precision Not high, power consumption etc..Using the photographic head of mobile terminal, whether photographic head is photographed with vehicle to judge the whether safe skill of pedestrian Art, does not still possess accuracy rate high and leave your defect too short to the time of user after sending alarm for.
Content of the invention
The technical problem to be solved is, provides a kind of " walking of diverting one's attention " early warning system based on mobile terminal System, it is possible to achieve safe early warning.
For solving above-mentioned technical problem, the present invention provides a kind of " walking of diverting one's attention " early warning system based on mobile terminal, bag Include model training module and real-time detection module;Model training module is by carrying out pre- place to a large amount of roadway characteristic training datasets Reason and model training, generate sample pattern and are used for real-time detection;Real-time detection module carries out sampling and pre- place using mobile terminal Reason, for obvious road markings feature, carries out classification according to the sample pattern that model training module generates to the sample of collection pre- Survey, decision-making judgement is carried out according to the result of classification prediction, when the user judging to divert one's attention walks out footpath, user is carried out pre- Alert.
Preferably, model training module includes pretreatment unit, model training unit and sample pattern, and pretreatment unit is used In a large amount of training datasets are carried out with pretreatment, the data that model training unit is used for pretreatment unit is processed carries out model instruction Practice thus producing sample pattern;Real-time detection model includes sample collection unit, pretreatment unit, classification predicting unit, decision-making Unit and user's prewarning unit, sample collection unit is used for sampling, and pretreatment unit is used for carrying out pretreatment to the sample of collection, Classification predicting unit is used for, according to the sample pattern that model training module generates, the sample of collection is carried out with classification prediction, decision-making list Unit carries out decision-making judgement thus carrying out early warning by user's prewarning unit to user for the result that classification is predicted.
Preferably, the target of system detectio is sidewalk for visually impaired people, and training dataset simulation pedestrian go out, and is completed by post-positioned pick-up head The collection of data, there is sidewalk for visually impaired people path and there is not sidewalk for visually impaired people path in image in the data of collection in image.
Preferably, the pretreatment unit in model training module carries out pretreatment to a large amount of training datasets, including ashing It is considered to multiple parameter values, the model that different parameters are generated is tested for image and adjustment of image size, select run time and The all preferable model of accuracy rate, records corresponding parameter value.
Preferably, the post-positioned pick-up head of the mobile terminal that sample collection unit is used based on user, photographic head is transported on backstage Under the control of mobile phone application in row, each cycle of operation start start and obtain the image in now photographic head, will Image sends the pretreatment unit in real-time detection module to.
Preferably, in each cycle of operation, the acceptance of decision-making judging unit is single to predict the outcome, and determines receiving this in advance After surveying result, provide whether user walks the judgement in the footpath with sidewalk for visually impaired people;Made using predicting the outcome of multiple image Judge, predicting the outcome and the front record predicting the outcome several times using this cycle of operation;The result of decision-making is 3 kinds:The positive, user It is in the footpath with sidewalk for visually impaired people, enter next cycle of operation after time interval T;Feminine gender, user is not in sidewalk for visually impaired people In footpath, enter next cycle of operation after time interval T;No, do not provide user's current state, directly sample again, Enter next cycle of operation.
Preferably, the step of decision-making is:
(1) set the value of Treq, the usual value of Treq 1~3, and to initialize the value of intermediate variable Tdet be zero;
(2) this cycle of operation predict the outcome as Pnow, if knot has had been made in a upper cycle of operation The judgement of fruit, and judged result Rprv is identical with Pnow, then the decision-making of this cycle of operation remains as Rprv, in time interval The step (2) of next cycle of operation is entered after T;
(3) if Rprv is different from Pnow, then judge the Pprv whether phase that predicts the outcome of Pnow and previous cycle of operation With;If identical, then make the value of Tdet increase by 1, and enter next step;If it is different, then the value of Tdet is entered as zero, when secondary The result of decision be no, then immediately proceed to the step (2) of next cycle of operation;
(4) judge whether the value of Tdet is not less than the value of Treq, if so, then the result of decision of this cycle of operation is Pnow;If it is not, when the secondary result of decision is no, then entering the step (2) of next cycle of operation after time interval T.
Beneficial effects of the present invention are:When user's handheld mobile phone moves towards driveway from footpath, oblique due to holding mobile phone Under angle, and hold the presence of mobile phone height, the picture of mobile phone camera capture is by for the road surface in front of user, the system energy Enough just provide warning before the deathtrap of user aisle in advance;On the basis of customizing without additional hardware, can be to " step of diverting one's attention Mobile terminal user OK " carries out effective early warning, and front testing result several times can be carried out with effectively utilizes, and accuracy is high, consumption Can be low, can be judged in advance.
Brief description
Fig. 1 is the overall system structure schematic diagram of the present invention.
Fig. 2 is the decision method schematic flow sheet of the present invention.
Fig. 3 is early warning scene graph during user's use of the present invention.
Specific embodiment
As shown in figure 1, a kind of " walking of diverting one's attention " early warning system based on mobile terminal, including model training module with real time Detection module;Model training module, by a large amount of training datasets are carried out with pretreatment and model training, generates sample pattern and uses In real-time detection;Real-time detection module is sampled using mobile terminal and pretreatment, the sample being generated according to model training module This model carries out classification prediction to the sample of collection, carries out decision-making judgement according to the result of classification prediction, the use divert one's attention when judgement When footpath is walked out at family, early warning is carried out to user.
The target of system detectio is sidewalk for visually impaired people, and when training dataset is gone out by simulating pedestrian, walking is while use handss Machine, completes by post-positioned pick-up head to gather.The data of collection is classified into two classes, and a class is to there is sidewalk for visually impaired people path in image, One class is to there is not sidewalk for visually impaired people.Two kinds of data correspond to safety zone and the insecure area judgement of decision-making judged result, once Correct decision-making judges that the image that can there will be sidewalk for visually impaired people is judged as safety zone, and the judgement that there will be no sidewalk for visually impaired people path is non-peace Region-wide.
Training dataset is carried out pretreatment by model training module first, including being ashed image and adjustment of image size, so It is trained setting up model with pretreated data afterwards.In this course it will be considered that multiple parameter values, including adjusting image Size, classifier type of selection etc..Then to different parameters generate model test, select run time and The all comparatively ideal model of accuracy rate, notes down corresponding parameter value, for the real-time detection on mobile terminal.By image procossing With dynamic sampling characteristic, improve the corresponding real-time of system;By Image semantic classification, including ashing image and adjustment of image size, From the sorter model that run time and accuracy rate are all more satisfactory, improve response speed.
After collecting sufficient amount training dataset image, pre- by image is carried out according to the parameter of setting to training data Process, process step includes being ashed image and adjustment of image size, being provided with parameter is picture size size.Then model instruction Practicing unit will be based on pretreated training dataset image, and the parameter according to setting carries out model training, sets up for real-time The sample pattern of the classification predicting unit of detection module, the parameter being provided with includes selecting in the classifier type adopting and image Feature point number.Set up after sample pattern according to these parameters, model training unit also will be carried out to the model establishing Adjust ginseng work, set up after different sample patterns using part training sample according to different parameter value collocation at random, so Using this model, other untapped training datasets are carried out with test prediction afterwards, veritify and predict the outcome and legitimate reading whether Show and calculate predictablity rate, record the time-consuming of this prediction process simultaneously.Count the accuracy rate of different sample patterns with Time-consuming result, then from going out run time and all comparatively ideal sample pattern of accuracy rate, using this model corresponding image chi The parameter value of very little size, classifier type and feature point number is for setting up the final sample pattern using.
The post-positioned pick-up head of the mobile terminal that the sample collection unit in real-time detection module is used based on user.Photographic head By in running background mobile phone application control under, each cycle of operation start most start and capture now photographic head In picture, image is sent to pretreatment unit.
The single image that pretreatment unit in each cycle of operation, real-time detection module is obtained carries out pre- place Reason, preprocess method is consistent with preprocess method in model training module.
Classification predicting unit in each cycle of operation, real-time detection module will be selected according to from model training module Model and parameter, come to this unit obtain pretreated image carry out classification prediction, to judge whether comprise in image Sidewalk for visually impaired people, predicts the outcome and will send to decision package.
As shown in Fig. 2 the decision package in each cycle of operation, real-time detection module predicts the outcome single for acceptance, And determine receiving after this predicts the outcome, the judgement of current environment can be provided, that is, whether user walks in the People's Bank of China with sidewalk for visually impaired people Judgement on road.Decision thought is, the prediction not entirely accurate of single image, but if the prediction knot using multiple image Fruit judging, the record predicting the outcome several times before using in other words, then the judging nicety rate made will greatly promote. The foundation of decision-making predicting the outcome and the front record predicting the outcome several times for this cycle of operation.The result of decision-making is 3 kinds:Positive (user is in the footpath with sidewalk for visually impaired people, enters next cycle of operation after time interval T), negative (user is not in carrying In the footpath of sidewalk for visually impaired people, enter next cycle of operation after time interval T), no (do not provide user's current state, directly again Secondary sampling, enters next cycle of operation).The step of decision-making is:
(1) set the value (Treq value is usually 1~3) of Treq, its meaning is to consider this basis predicting the outcome On, this decision-making also to be considered before the predicting the outcome of Treq cycle of operation.And initialize intermediate variable Tdet's Value is zero, and its meaning is to judge whether the number that predicts the outcome currently considered meets Treq and require.
(2) this cycle of operation predict the outcome as Pnow, if knot has had been made in a upper cycle of operation The judgement of fruit, and judged result Rprv is identical with Pnow, then the decision-making of this cycle of operation remains as Rprv, in time interval The step (2) of next cycle of operation is entered after T.
(3) if Rprv is different from Pnow, then judge the Pprv whether phase that predicts the outcome of Pnow and previous cycle of operation With.If identical, then make the value of Tdet increase by 1, and enter next step;If it is different, then the value of Tdet is entered as zero, when secondary The result of decision be no, then immediately proceed to the step (2) of next cycle of operation.
(4) judge whether the value of Tdet is not less than the value of Treq, if so, then the result of decision of this cycle of operation is Pnow;If it is not, when the secondary result of decision is no, then entering the step (2) of next cycle of operation after time interval T.
As shown in figure 3, when user's handheld mobile phone moves towards driveway from footpath, due to holding mobile phone angle obliquely, And hold the presence of mobile phone height, the picture of mobile phone camera capture is by for the road surface in front of user.Also because this feature, this System just can provide warning before user goes to deathtrap in advance.
The present invention adopts dynamic sampling interval and photaesthesia feature, mitigates system energy consumption expense, and reserved user is more tight The response time of anxious situation.System will be according to user's run trace, if user goes on along in safety zone, system will suitably increase Plus the sampling interval, reduce system energy consumption, but constrain the maximum sampling interval simultaneously, it is to avoid the sampling interval infinitely expands, and reduces system Safety.If user passes through often walks in dangerous region, system suitably will reduce the sampling interval, enable a system to preferably Find dangerous situation, give user warning.In the scene that night or light are bad, if user uses mobile phone by extreme influence row Road safety, and system nor judge safety, will quit work and persistently report to the police, this operation can reduce system energy consumption with When ensure user security.And be also the mistake in an important dynamic regulation sampling interval of the system in above-mentioned decision method Journey, when the detection of system has determined different non-footpath and judges from footpath before, system will be skipped the sampling interval, Directly carry out sampling next time and detection.This is due to coming from two kinds of probabilities, detection mistake and user the occurrence of this Walk out footpath it is therefore desirable to detect to draw correct judgement immediately again, if it is determined that being that user walks out footpath, then will Provide warning, the method just can keep for user's more time to tackle emergency.
Roadway characteristic is detected by mobile terminal camera, there is wide applicability.In view of current every country is all adopted The mode arranging label information used in road reminds pedestrian, and for example Belgium's white line on road forms and " sends short messages special Road ".The present invention by only changing training dataset, and can not change the mode of any software, is applied to new scene, tool There is wide applicability.
The system, when changing detection object, is applied to transfer real-time detection module during scene beyond sidewalk for visually impaired people is detected, Only need to change the training dataset of model training module, using identical systems unit and method, image is carried out to data set and locate in advance Manage, set up sample pattern and adjust ginseng, you can obtain the final sample model corresponding to this detection object, for real-time detection module Use.And the difference degree that the accuracy rate of generally detection is concentrated inhomogeneity sample by training data is affected.In the system In test, except the detection of sidewalk for visually impaired people, the detection of zebra crossing, and the detection arranging visible marking in road also have excellent standard Really rate and detection time performance.
Although the present invention is illustrated with regard to preferred implementation and has been described, it is understood by those skilled in the art that Without departing from scope defined by the claims of the present invention, variations and modifications can be carried out to the present invention.

Claims (7)

1. a kind of " walking of diverting one's attention " early warning system based on mobile terminal is it is characterised in that include:Model training module is with real time Detection module;Model training module, by a large amount of roadway characteristic training datasets are carried out with pretreatment and model training, generates sample This model is used for real-time detection;Real-time detection module is sampled using mobile terminal and pretreatment, for obvious road markings Feature, carries out classification prediction according to the sample pattern that model training module generates to the sample of collection, according to the knot of classification prediction Fruit carries out decision-making judgement, when the user judging to divert one's attention walks out footpath, carries out early warning to user.
2. " walking of the diverting one's attention " early warning system based on mobile terminal as claimed in claim 1 is it is characterised in that model training mould Block includes pretreatment unit, model training unit and sample pattern, and pretreatment unit is pre- for carrying out to a large amount of training datasets Process, the data that model training unit is used for pretreatment unit is processed carries out model training thus producing sample pattern;In real time Detection model includes sample collection unit, pretreatment unit, classification predicting unit, decision package and user's prewarning unit, sample Collecting unit is used for sampling, and pretreatment unit is used for the sample of collection is carried out pretreatment, and classification predicting unit is used for according to mould The sample pattern that type training module generates carries out classification prediction to the sample of collection, and decision package is used for the result to classification prediction Carry out decision-making judgement thus early warning is carried out to user by user's prewarning unit.
3. as claimed in claim 1 " walking of the diverting one's attention " early warning system based on mobile terminal is it is characterised in that system detectio Target is sidewalk for visually impaired people or other roads mark feature, and training dataset simulation pedestrian go out, and completes data by post-positioned pick-up head Collection, there is sidewalk for visually impaired people path and there is not sidewalk for visually impaired people path in image in the data of collection in image.
4. " walking of the diverting one's attention " early warning system based on mobile terminal as claimed in claim 2 is it is characterised in that model training mould Pretreatment unit in block carries out pretreatment to a large amount of training datasets, including ashing image and adjustment of image size it is considered to many Individual parameter value, the model that different parameters are generated is tested, and selects run time and all preferable model of accuracy rate, it is right to record The parameter value answered.
5. " walking of the diverting one's attention " early warning system based on mobile terminal as claimed in claim 2 is it is characterised in that sample collection list Unit based on user use mobile terminal post-positioned pick-up head, photographic head in running background mobile phone application control under, The starting of each cycle of operation starts and obtains the image in now photographic head, and image is sent in real-time detection module Pretreatment unit.
6. " walking of the diverting one's attention " early warning system based on mobile terminal as claimed in claim 2 is it is characterised in that run at each Cycle, decision-making judging unit accept single predicts the outcome, and determine receiving after this predicts the outcome, provide whether user walks Judgement in footpath with sidewalk for visually impaired people;Judged using predicting the outcome of multiple image, using the prediction of this cycle of operation Result and the front record predicting the outcome several times;The result of decision-making is 3 kinds:The positive, user is in the footpath with sidewalk for visually impaired people, when Between enter next cycle of operation after the T of interval;Feminine gender, user is not in the footpath with sidewalk for visually impaired people, enters after time interval T Next cycle of operation;No, do not provide user's current state, directly sample again, enter next cycle of operation.
7. " walking of the diverting one's attention " early warning system based on mobile terminal as claimed in claim 6 is it is characterised in that the step of decision-making For:
(1) set the value of Treq, the usual value of Treq 1~3, and to initialize the value of intermediate variable Tdet be zero;
(2) this cycle of operation predict the outcome as Pnow, if having been made resultful in a upper cycle of operation Judge, and judged result Rprv is identical with Pnow, then the decision-making of this cycle of operation remains as Rprv, after time interval T Enter the step (2) of next cycle of operation;
(3) if Rprv is different from Pnow, then judge whether Pnow is identical with the Pprv that predicts the outcome of previous cycle of operation;If Identical, then to make the value of Tdet increase by 1, and enter next step;If it is different, then the value of Tdet is entered as zero, when secondary is determined Plan result is no, then immediately proceeds to the step (2) of next cycle of operation;
(4) judge whether the value of Tdet is not less than the value of Treq, if so, then the result of decision of this cycle of operation is Pnow;If It is not when the secondary result of decision is no, then to enter the step (2) of next cycle of operation after time interval T.
CN201610881571.XA 2016-10-09 2016-10-09 A kind of " walking of diverting one's attention " early warning system based on mobile terminal Active CN106408932B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610881571.XA CN106408932B (en) 2016-10-09 2016-10-09 A kind of " walking of diverting one's attention " early warning system based on mobile terminal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610881571.XA CN106408932B (en) 2016-10-09 2016-10-09 A kind of " walking of diverting one's attention " early warning system based on mobile terminal

Publications (2)

Publication Number Publication Date
CN106408932A true CN106408932A (en) 2017-02-15
CN106408932B CN106408932B (en) 2018-11-16

Family

ID=59228338

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610881571.XA Active CN106408932B (en) 2016-10-09 2016-10-09 A kind of " walking of diverting one's attention " early warning system based on mobile terminal

Country Status (1)

Country Link
CN (1) CN106408932B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108524209A (en) * 2018-03-30 2018-09-14 江西科技师范大学 Blind-guiding method, system, readable storage medium storing program for executing and mobile terminal
CN108960029A (en) * 2018-03-23 2018-12-07 北京交通大学 A kind of pedestrian diverts one's attention behavioral value method
CN109241916A (en) * 2018-09-12 2019-01-18 四川长虹电器股份有限公司 A kind of system and method for pedestrian's walking safety detection based on smart phone

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080309468A1 (en) * 2007-06-12 2008-12-18 Greene Daniel H Human-machine-interface (HMI) customization based on collision assessments
CN101853399A (en) * 2010-05-11 2010-10-06 北京航空航天大学 Method for realizing blind road and pedestrian crossing real-time detection by utilizing computer vision technology
CN101964054A (en) * 2010-09-29 2011-02-02 东南大学 Friendly track detection system based on visual processing
CN102609724A (en) * 2012-02-16 2012-07-25 北京航空航天大学 Method for prompting ambient environment information by using two cameras
CN103908365A (en) * 2014-04-09 2014-07-09 天津思博科科技发展有限公司 Electronic travel assisting device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080309468A1 (en) * 2007-06-12 2008-12-18 Greene Daniel H Human-machine-interface (HMI) customization based on collision assessments
CN101853399A (en) * 2010-05-11 2010-10-06 北京航空航天大学 Method for realizing blind road and pedestrian crossing real-time detection by utilizing computer vision technology
CN101964054A (en) * 2010-09-29 2011-02-02 东南大学 Friendly track detection system based on visual processing
CN102609724A (en) * 2012-02-16 2012-07-25 北京航空航天大学 Method for prompting ambient environment information by using two cameras
CN103908365A (en) * 2014-04-09 2014-07-09 天津思博科科技发展有限公司 Electronic travel assisting device

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108960029A (en) * 2018-03-23 2018-12-07 北京交通大学 A kind of pedestrian diverts one's attention behavioral value method
CN108524209A (en) * 2018-03-30 2018-09-14 江西科技师范大学 Blind-guiding method, system, readable storage medium storing program for executing and mobile terminal
CN109241916A (en) * 2018-09-12 2019-01-18 四川长虹电器股份有限公司 A kind of system and method for pedestrian's walking safety detection based on smart phone

Also Published As

Publication number Publication date
CN106408932B (en) 2018-11-16

Similar Documents

Publication Publication Date Title
CN112069944B (en) Road congestion level determining method
CN104933870B (en) Vehicle fake plate identification method and device based on vehicle behavior analysis
CN109190488B (en) Front vehicle door opening detection method and device based on deep learning YOLOv3 algorithm
CN111563494A (en) Behavior identification method and device based on target detection and computer equipment
CN108509954A (en) A kind of more car plate dynamic identifying methods of real-time traffic scene
CN107170247B (en) Method and device for determining queuing length of intersection
CN109495654B (en) Pedestrian safety sensing method based on smart phone
KR102015947B1 (en) Method for extracting image of learning object for autonomous driving and apparatus thereof
CN111460924B (en) Gate ticket-evading behavior detection method based on target detection
CN105539450B (en) A kind of automatic identifying method and device driving stroke
CN103324955A (en) Pedestrian detection method based on video processing
CN110866479A (en) Method, device and system for detecting that motorcycle driver does not wear helmet
CN104123532B (en) Target object detection and target object quantity confirming method and device
CN103909826A (en) Optimization method for collaboratively sensing violation behavior of drivers
CN106408932A (en) Mobile terminal based pre-warning system for distractive walks
CN104112279B (en) A kind of object detection method and device
CN110322687B (en) Method and device for determining running state information of target intersection
CN103345842A (en) Road vehicle classification system and method
CN106682600A (en) Method and terminal for detecting targets
CN110674887A (en) End-to-end road congestion detection algorithm based on video classification
CN103324957A (en) Identification method and identification device of state of signal lamps
CN114494998B (en) Intelligent analysis method and system for vehicle data
CN115223124A (en) Method and device for fitting lane line equation, vehicle and storage medium
CN113887431A (en) AI-based detection method for identifying person without gloves in kitchen scene
TW201822168A (en) Vehicle moving direction predicting system and method using digital image recognition in combination with moving trace computation technology with application of big data computation technology

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant