CN103810854A - Intelligent traffic algorithm evaluation method based on manual calibration - Google Patents
Intelligent traffic algorithm evaluation method based on manual calibration Download PDFInfo
- Publication number
- CN103810854A CN103810854A CN201410074978.2A CN201410074978A CN103810854A CN 103810854 A CN103810854 A CN 103810854A CN 201410074978 A CN201410074978 A CN 201410074978A CN 103810854 A CN103810854 A CN 103810854A
- Authority
- CN
- China
- Prior art keywords
- traffic
- vehicle
- video
- virtual coil
- track
- 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
Images
Landscapes
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to an intelligent traffic algorithm evaluation method based on manual calibration. The method includes that a traffic video database is established; the manual calibration is performed on traffic video by ViPER software; calibrated parameters are saved as a document in XML format and guided out; data analysis is performed on the XML document to obtain required traffic parameters; the traffic video database is improved; algorithms to be verified are evaluated. Compared with traffic parameters intelligent algorithms based on computer vision, the intelligent traffic algorithm evaluation method based on the manual calibration has the advantages that time-location information of vehicle/pedestrian/traffic signs in the traffic video is manually recognized and calibrated, the traffic parameters are automatically obtained by a computer, the accuracy of the traffic parameters is greatly improved, an evaluation platform of multiple traffic parameter intelligent algorithms is established, the algorithms to be verified can be subjected to multi-scene, multi-parameter and all-round verification and evaluation, the evaluation method is accurate and reliable, and a certain guide effect on the improvement of the evaluated algorithms can be achieved.
Description
Technical field
The invention belongs to machine vision and field of intelligent control, is a kind of method of utilizing computer technology, image processing techniques, database technology etc. to assess intelligent transportation algorithm.
Background technology
Along with improving gradually and the increase of private car quantity of highway traffic infrastructure, road traffic problem is day by day serious.In recent years, frequent accidents, the Loss of Life and property that road traffic accident causes more and more causes people's concern.According to statistics: 2008,265204 of road traffic accidents occurred altogether in the whole nation, caused 73484 people's death, and 304919 people are injured, 10.1 hundred million yuan of direct property losss; 2009, there is 238351 of road traffic accidents in the whole nation altogether, causes 67759 people's death, and 275125 people are injured, 9.1 hundred million yuan of direct property losss; 2010, there is 3906164 of road traffic accidents in the whole nation altogether, causes 65225 people's death, and 254075 people are injured, 9.3 hundred million yuan of direct property losss.The Loss of Life and property causing in order to reduce to greatest extent road traffic accident, in the past few decades, expert and scholar are in the research work of actively developing about intelligent transportation system (ITS) both at home and abroad.
In recent years, numerous intelligent transportation algorithms based on computer vision arise at the historic moment, and they use the technology such as computing machine and image processing, complete the automatic detection to comprising the traffic parameters such as vehicle number, the speed of a motor vehicle, vehicle.Algorithm difference, precision is also different.Although the accuracy rate that various intelligent algorithms are announced is mostly more than 90%, but owing to going back up to now unified, accurate, the reliable verification platform of neither one and appraisal procedure, whether the accuracy that cannot verify various algorithms announcements is credible, also with regard to being difficult to, the quality of various algorithms compared to judge.
Summary of the invention
For the problems referred to above, the present invention proposes a kind of intelligent transportation algorithm evaluation method based on artificial demarcation, based on the data of artificial demarcation, obtain traffic parameter accurately, and the traffic parameter that itself and intelligent algorithm to be assessed obtain is contrasted, obtain the accuracy of this algorithm, thereby evaluate the quality of this algorithm performance.
Hardware device involved in the present invention comprises PC, video camera and corresponding capture card.The video that video camera is taken imports PC, and is converted into the video file of general format, as MPG form.Manually demarcate part and need application open source software ViPER, this software can mark target unit residing position in video manually frame by frame, and the present invention applies this software and carries out time-position data demarcation of traffic unit's (vehicle/pedestrian/traffic sign).
An intelligent transportation algorithm evaluation method based on artificial demarcation, comprises the following steps:
Step 1, sets up traffic video database.
Collect the traffic video under many scenes, varying environment, visual classification is as follows:
Different location: through street, intersection, high speed crossing ring road etc.;
The different periods: morning peak, evening peak, daytime, at dusk etc.;
Varying environment: shadow-free, illumination shade is obvious, mist, sleety weather etc.;
Different target: vehicle, pedestrian, traffic sign etc.;
Every section of duration 10～60 minutes not etc., can not meet the checking demand of many algorithms.
Step 2, utilizes ViPER software manually to demarcate traffic video.
Step 2.1, is written into traffic video to be calibrated.
Step 2.2, the parameter that setting will be demarcated, comprises title, callout box kind etc.
Step 2.3, the time-positional information of manually demarcating each unit in traffic video.
Artificial information of demarcating comprises: unit numbering; Classification: 1-vehicle, 2-pedestrian, 3-traffic sign; Unit lasting frame number in video; The upper left corner point coordinate (X, Y) of unit area; The transverse width Width of unit area, longitudinal height H eight.
Step 3, derives artificial data of demarcating.
The data of demarcating are saved as to the file of XML form in ViPER software, and by its derivation.
Step 4, carries out data analysis to XML file described in step 3, obtains required traffic parameter.
Step 4.1, forms vehicle travel track.
Read demarcation information, the rectangle frame center (X+Width/2, Y+Height/2) being made up of X, Y, Width, Height is considered as to vehicle position, link the vehicle location of each time point, form vehicle travel track.
Step 4.2, calculates vehicle number.
(1) set number of track-lines and virtual coil according to video, coil width and track are with wide, and length is not less than a length of wagon.
(2), according to the geometric relationship of vehicle and virtual coil, carry out divided lane vehicle count.
As shown in Figure 1, quadrilateral ABCD is the virtual coil of setting, and vehicle V sails CD limit into by AB limit and rolls away from.There is geometric relationship VA+VB-AB >=0, VC+VD-CD >=0 in four summits of vehicle location and virtual coil.The corresponding vehicle of minimum point of VA+VB-AB enters the moment of this virtual coil, in the moment that the corresponding vehicle of minimum point of VC+VD-CD rolls virtual coil away from, can judge accordingly the moment that vehicle sailed or rolled away from the behavior of coil into and enters and roll away from.
If certain car completes to same virtual coil the behavior of rolling away from of sailing into, carry out the once vehicle count to this track.This vehicle number+1, track.
If certain vehicle sails certain virtual coil into, but do not roll any virtual coil away from, think that this car travels extremely, does not count.
If certain vehicle sails certain virtual coil into, but roll away from from another virtual coil, think that this car changes behavior, the virtual coil counting+1 that it rolls away from.
Step 4.3, calculates following distance, average following distance.
According to camera calibration, obtain the corresponding physical length of road in video.Subtracted each other and obtained following distance by adjacent rectangle frame coordinate position, average following distance is the following distance average in statistics a period of time.
Step 4.4, calculates the speed of a motor vehicle, average speed.
According to camera calibration, obtain the corresponding actual range of drawn virtual coil, in the moment that obtains each car and enter and roll away from virtual coil according to method described in step 4.2, obtain Vehicle Speed by the distance/time.In statistics a period of time, the speed of all vehicles through virtual coil, obtains its average velocity.
Step 4.5, calculates vehicle queue length.
Vehicle queue is a kind of traffic ripple, therefore queue length is different from the existence judgement of vehicle, not only relevant to current location, the velocity information of each car in scene, also relevant to its historical information, only in a certain frame from single image or video, can not obtain queue length accurately.The artificial calibrated and calculated vehicle queue length that the present invention is based on multiframe, long-time section, concrete grammar is as follows:
Set stop line position, utilize the time-positional information of each car that step 1 obtains to draw its relation curve apart from 'STOP' line ahead and time, as shown in Figure 2, a car of every line correspondence.Arrive the queuing vehicle of intersection through travel-parking waiting-continue to travel 3 stages, stop and be called flex point with the separation starting.The flex point of each car in matching track obtains the parking ripple A (t) and startup ripple B (t) in whole track, i.e. traffic ripple, as shown in Figure 3, the corresponding stop line of initial point position.Calculate queue length and stop delay according to the relation of traffic ripple and queue length:
QL(t)=TL(t)-HL(t)
In formula, t is the time, and QL (t) is queue length, and TL (t) is tail of the queue and 'STOP' line ahead, and HL (t) is enemy and 'STOP' line ahead.
Step 4.6, obtains traffic sign positional information.
Read the road signs information of Vehicular video mark, obtain the position of traffic sign in each moment, pass through camera calibration, simulate the length relation in actual range correspondence image, be converted into actual range information by the positional information of traffic sign, show in real time the distance between this traffic sign and driving vehicle.
Step 4.7, forms pedestrian's track according to the method for step 4.1.
Step 5, improves traffic video database.
Every section of traffic video in step 1 is manually demarcated according to method described in step 2 and step 3, then calculate the traffic parameter in every section of video according to method described in step 4, original video and corresponding artificial nominal data and the common traffic video database forming for algorithm evaluation of traffic parameter thereof.
Step 6, treats verification algorithm and assesses.
In selecting video database more than same section or section video, apply the above-mentioned traffic parameter of algorithm measurement to be verified, the traffic parameter in measurement result and database is compared, the accuracy of computation and measurement, judges the accuracy of algorithm to be verified thus.
Compared with prior art, the present invention has the following advantages:
1. the present invention is by the time-positional information of the vehicle/pedestrian/traffic sign in artificial cognition and demarcation traffic video, by computing machine automatic acquisition traffic parameter, compared with traffic parameter intelligent algorithm based on computer vision, greatly improve the precision that traffic parameter calculates; Compared with the method for pure manual record, efficiency is higher, and is easy to storage, the error that also can avoid some artificial processing eases to occur.
2. the database that the present invention application comprises traffic video and artificial nominal data and traffic parameter, can carry out many scenes, multiparameter, checking assessment in all directions to traffic parameter intelligent algorithm to be verified.Owing to having used high precision traffic parameter, guarantee accuracy and the reliability of checking assessment.Appraisal procedure of the present invention has certain directive function to the Improvement and perfection of evaluated algorithm.
Accompanying drawing explanation
Fig. 1 is virtual coil method of counting schematic diagram;
Fig. 2 is the relation curve of queuing vehicle apart from 'STOP' line ahead and time;
Fig. 3 is that traffic ripple and queue length are related to schematic diagram;
Fig. 4 is the process flow diagram of method involved in the present invention;
Fig. 5 is vehicle time-positional information schematic diagram of artificial mark in video;
Fig. 6 is with the video schematic diagram of markup information and virtual coil.
Embodiment
Below in conjunction with drawings and Examples, the present invention will be further described.
Application the present invention assesses a kind of image processing algorithm of divided lane counting.Embodiment realizes on the PC that VC2008 and OpenCV2.3.1 are installed.
The process flow diagram of appraisal procedure of the present invention as shown in Figure 4, comprises the following steps:
Step 1, sets up traffic video database.
Database at least comprises the traffic video of 4 kinds of varying environments, and file layout is MPG.
Step 2, utilizes ViPER software manually to demarcate traffic video.
Step 2.1, is written into one section of vehicle by the traffic video yinying.mpg of solar radiation generation shade.
Step 2.2, the parameter that setting will be demarcated, the vehicle title of mark is set as Vehicle, callout box kind is Bbox(rectangle frame).
Step 2.3, manually marks out residing position of each moment of vehicle.As shown in Figure 5.
Step 3, derives artificial data of demarcating.
The parameter of demarcating in ViPER software saves as the file of XML form, by its preservation derivation.Export name is called yinying-mpg.xml.In file, preserve the time-positional information of the unit of all artificial demarcation
Take a certain car as example, XML data file comprises following parameter:
Car number ID=0;
Classification KIND=1(vehicle);
Vehicle lasting frame number framespan=1:45 in video;
Detailed data Data: comprise the upper left corner point coordinate of each two field picture vehicle region, width and the height of unit area.
Step 4, carries out data analysis by the XML file of the time-positional information that has unit in video, to obtain required traffic parameter---and vehicle number (divided lane).
Step 4.1, setting number of track-lines according to video is 4, and sets virtual coil, and coil width and track are with wide, and length is not less than a length of wagon.As shown in Figure 6.
Step 4.2, reads XML file and obtains divided lane count results.
Four track count results are respectively 3,16,27,7.
Step 5, is written into other video-frequency band, manually demarcates respectively and calculation of parameter according to method described in step 2～5, improves original video data storehouse.Wherein, the divided lane of traffic video xue.mpg of one section of snow day is counted as 10,22, and 26,9, one section is counted as 18,30,39,32 without the divided lane of the traffic video normal.mpg of any environmental interference clearly substantially.
Step 6 is moved algorithm to be verified on above-mentioned 3 sections of identical videos, and obtains four track count results.Video yinying.mpg is 3,15,27,6; Video xue.mpg is 3,11,10,3; Video normal.mpg is 18,30,40,32.The traffic parameter that the true traffic parameter of every section of video and algorithm to be verified are recorded contrasts, and assesses algorithm accuracy to be verified.
For video yinying.mpg, the processing accuracy evaluation result of algorithm to be verified is 94.87%, substantially meets accuracy requirement.
For video xue.mpg, the processing accuracy evaluation result of algorithm to be verified is 37.95%, cannot meet accuracy requirement.
For video normal.mpg, the processing accuracy evaluation result of algorithm to be verified is 99.38%, meets accuracy requirement.
Can draw the following conclusions according to assessment result: algorithm to be verified is affected not quite by illumination and shade etc., and under other atrocious weather environment such as sleet, the precision of algorithm declines greatly, cannot meet accuracy requirement.
Claims (4)
1. the intelligent transportation algorithm evaluation method based on artificial demarcation, is characterized in that, comprises the following steps:
Step 1, sets up traffic video database;
Collect the traffic video under many scenes, varying environment, visual classification is as follows:
Different location: through street, intersection, high speed crossing ring road;
The different periods: morning peak, evening peak, daytime, at dusk;
Varying environment: shadow-free, illumination shade is obvious, mist, sleety weather;
Different target: vehicle, pedestrian, traffic sign;
Every section of duration 10～60 minutes not etc., can not meet the checking demand of many algorithms;
Step 2, utilizes ViPER software manually to demarcate traffic video;
Step 2.1, is written into traffic video to be calibrated;
Step 2.2, the data that setting will be demarcated, comprise title, callout box kind;
Step 2.3, each unit of manually demarcating in traffic video is the time-positional information of vehicle/pedestrian/traffic sign;
Step 3, derives artificial data of demarcating;
The data of demarcating are saved as to the file of XML form in ViPER software, and by its derivation;
Step 4, carries out data analysis to XML file described in step 3, obtains required traffic parameter;
Step 4.1, forms vehicle travel track;
Read demarcation information, the rectangle frame center (X+Width/2, Y+Height/2) being made up of X, Y, Width, Height is considered as to vehicle position, link the vehicle location of each time point, form vehicle travel track;
Step 4.2, calculates vehicle number;
Step 4.3, calculates following distance, average following distance;
According to camera calibration, obtain the corresponding physical length of road in video; Subtracted each other and obtained following distance by adjacent rectangle frame coordinate position, average following distance is the following distance average in statistics a period of time;
Step 4.4, calculates the speed of a motor vehicle, average speed;
According to camera calibration, obtain the corresponding actual range of drawn virtual coil, in the moment that obtains each car and enter and roll away from virtual coil according to method described in step 4.2, obtain Vehicle Speed by the distance/time; In statistics a period of time, the speed of all vehicles through virtual coil, obtains its average velocity;
Step 4.5, calculates vehicle queue length;
Step 4.6, obtains traffic sign positional information;
The road signs information that reads Vehicular video mark, obtains traffic sign in the position in each moment, by camera calibration, simulates the length relation in actual range correspondence image, is converted into actual range information by the positional information of traffic sign;
Step 4.7, forms pedestrian's track according to the method for step 4.1;
Step 5, improves traffic video database;
Every section of traffic video described in step 1 is manually demarcated according to method described in step 2 and step 3, then calculate the traffic parameter in every section of video according to method described in step 4, original video and corresponding artificial nominal data and the common traffic video database forming for algorithm evaluation of traffic parameter thereof;
Step 6, treats verification algorithm and assesses;
In selecting video database more than same section or section video, apply the above-mentioned traffic parameter of algorithm measurement to be verified, the traffic parameter in measurement result and database is compared, the accuracy of computation and measurement, judges the accuracy of algorithm to be verified thus.
2. a kind of intelligent transportation algorithm evaluation method based on artificial demarcation according to claim 1, is characterized in that, the artificial information of demarcating of described step 2.3 comprises: unit numbering; Classification: 1-vehicle, 2-pedestrian, 3-traffic sign; Unit lasting frame number in video; The upper left corner point coordinate (X, Y) of unit area; The transverse width Width of unit area, longitudinal height H eight.
3. a kind of intelligent transportation algorithm evaluation method based on artificial demarcation according to claim 1, is characterized in that, it is further comprising the steps of that described step 4.2 is calculated the method for vehicle number:
(1) set number of track-lines and virtual coil according to video, coil width and track are with wide, and length is not less than a length of wagon;
(2), according to the geometric relationship of vehicle and virtual coil, carry out divided lane vehicle count;
Suppose that quadrilateral ABCD is the virtual coil of setting, vehicle V sails CD limit into by AB limit and rolls away from; There is geometric relationship VA+VB-AB >=0, VC+VD-CD >=0 in four summits of vehicle location and virtual coil; The corresponding vehicle of minimum point of VA+VB-AB enters the moment of this virtual coil, in the moment that the corresponding vehicle of minimum point of VC+VD-CD rolls virtual coil away from, can judge accordingly the moment that vehicle sailed or rolled away from the behavior of coil into and enters and roll away from;
If certain car completes to same virtual coil the behavior of rolling away from of sailing into, carry out the once vehicle count to this track; This vehicle number+1, track;
If certain vehicle sails certain virtual coil into, but do not roll any virtual coil away from, think that this car travels extremely, does not count;
If certain vehicle sails certain virtual coil into, but roll away from from another virtual coil, think that this car changes behavior, the virtual coil counting+1 that it rolls away from.
4. a kind of intelligent transportation algorithm evaluation method based on artificial demarcation according to claim 1, is characterized in that, the artificial calibrated and calculated vehicle queue length of described step 4.5 based on multiframe, long-time section, and concrete grammar is as follows:
Set stop line position, utilize the time-positional information of each car that step 1 obtains to draw its relation curve apart from 'STOP' line ahead and time, every line represents a car; Arrive the queuing vehicle of intersection through travel-parking waiting-continue to travel 3 stages, stop and be called flex point with the separation starting; The flex point of each car in matching track obtains the parking ripple and startup ripple, i.e. traffic ripple in whole track; Calculate queue length and stop delay according to the relation of traffic ripple and queue length:
QL(t)=TL(t)-HL(t)
In formula, t is the time, and QL (t) is queue length, and TL (t) is tail of the queue and 'STOP' line ahead, and HL (t) is enemy and 'STOP' line ahead.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410074978.2A CN103810854B (en) | 2014-03-03 | 2014-03-03 | A kind of based on the artificial intelligent transportation parameter detection method demarcated |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410074978.2A CN103810854B (en) | 2014-03-03 | 2014-03-03 | A kind of based on the artificial intelligent transportation parameter detection method demarcated |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103810854A true CN103810854A (en) | 2014-05-21 |
| CN103810854B CN103810854B (en) | 2016-04-06 |
Family
ID=50707562
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410074978.2A Active CN103810854B (en) | 2014-03-03 | 2014-03-03 | A kind of based on the artificial intelligent transportation parameter detection method demarcated |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103810854B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023436A (en) * | 2015-08-20 | 2015-11-04 | 南京安通杰科技实业有限公司 | Traffic information evaluation method |
| CN107301776A (en) * | 2016-10-09 | 2017-10-27 | 上海炬宏信息技术有限公司 | Track road conditions processing and dissemination method based on video detection technology |
| CN108447261A (en) * | 2018-04-04 | 2018-08-24 | 迈锐数据（北京）有限公司 | Based on multimode vehicle queue length computational methods and device |
| CN109522825A (en) * | 2018-10-31 | 2019-03-26 | 蔚来汽车有限公司 | The Performance Test System and its performance test methods of visual perception system |
| CN111141311A (en) * | 2019-12-31 | 2020-05-12 | 武汉中海庭数据技术有限公司 | Evaluation method and system of high-precision map positioning module |
| CN112257587A (en) * | 2020-10-22 | 2021-01-22 | 江苏禹空间科技有限公司 | Target object detection effect evaluation method and device, storage medium and equipment |
| CN113469109A (en) * | 2021-07-16 | 2021-10-01 | 阿波罗智联(北京)科技有限公司 | Traffic light identification result processing method and device, road side equipment and cloud control platform |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101145240A (en) * | 2007-11-09 | 2008-03-19 | 王海燕 | Camera image road multiple-point high precision calibration method |
| CN101917595A (en) * | 2010-07-01 | 2010-12-15 | 李志恒 | Video monitoring system for automatically marking road section information |
| CN102063789A (en) * | 2009-11-16 | 2011-05-18 | 高德软件有限公司 | Traffic information quality evaluation method and system |
| CN102222332A (en) * | 2011-05-19 | 2011-10-19 | 长安大学 | Geometric calibration method of camera under linear model |
| CN102402853A (en) * | 2011-12-08 | 2012-04-04 | 北京易华录信息技术股份有限公司 | Traffic information accuracy evaluation system and method |
| EP2602740A2 (en) * | 2011-12-08 | 2013-06-12 | Palo Alto Research Center Incorporated | Changing parameters of sequential video frames to detect different types of objects |
| KR20130066210A (en) * | 2011-12-12 | 2013-06-20 | 현대모비스 주식회사 | Device and method of acquiring traffic-control sign information using a camera |
-
2014
- 2014-03-03 CN CN201410074978.2A patent/CN103810854B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101145240A (en) * | 2007-11-09 | 2008-03-19 | 王海燕 | Camera image road multiple-point high precision calibration method |
| CN102063789A (en) * | 2009-11-16 | 2011-05-18 | 高德软件有限公司 | Traffic information quality evaluation method and system |
| CN101917595A (en) * | 2010-07-01 | 2010-12-15 | 李志恒 | Video monitoring system for automatically marking road section information |
| CN102222332A (en) * | 2011-05-19 | 2011-10-19 | 长安大学 | Geometric calibration method of camera under linear model |
| CN102402853A (en) * | 2011-12-08 | 2012-04-04 | 北京易华录信息技术股份有限公司 | Traffic information accuracy evaluation system and method |
| EP2602740A2 (en) * | 2011-12-08 | 2013-06-12 | Palo Alto Research Center Incorporated | Changing parameters of sequential video frames to detect different types of objects |
| KR20130066210A (en) * | 2011-12-12 | 2013-06-20 | 현대모비스 주식회사 | Device and method of acquiring traffic-control sign information using a camera |
Non-Patent Citations (2)
| Title |
|---|
| 李鹏飞 等: "智能视频算法评估综述", 《计算机辅助设计与图形学学报》 * |
| 杨德亮: "基于复式伸缩窗的车辆排队与消散快速检测算法", 《公路交通科技》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023436A (en) * | 2015-08-20 | 2015-11-04 | 南京安通杰科技实业有限公司 | Traffic information evaluation method |
| CN107301776A (en) * | 2016-10-09 | 2017-10-27 | 上海炬宏信息技术有限公司 | Track road conditions processing and dissemination method based on video detection technology |
| CN108447261A (en) * | 2018-04-04 | 2018-08-24 | 迈锐数据（北京）有限公司 | Based on multimode vehicle queue length computational methods and device |
| CN109522825A (en) * | 2018-10-31 | 2019-03-26 | 蔚来汽车有限公司 | The Performance Test System and its performance test methods of visual perception system |
| CN111141311A (en) * | 2019-12-31 | 2020-05-12 | 武汉中海庭数据技术有限公司 | Evaluation method and system of high-precision map positioning module |
| CN112257587A (en) * | 2020-10-22 | 2021-01-22 | 江苏禹空间科技有限公司 | Target object detection effect evaluation method and device, storage medium and equipment |
| CN112257587B (en) * | 2020-10-22 | 2023-12-22 | 无锡禹空间智能科技有限公司 | Target object detection effect evaluation method, device, storage medium and equipment |
| CN113469109A (en) * | 2021-07-16 | 2021-10-01 | 阿波罗智联(北京)科技有限公司 | Traffic light identification result processing method and device, road side equipment and cloud control platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103810854B (en) | 2016-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103810854B (en) | A kind of based on the artificial intelligent transportation parameter detection method demarcated | |
| CN102819954B (en) | Traffic region dynamic map monitoring and predicating system | |
| CN112700470B (en) | Target detection and track extraction method based on traffic video stream | |
| CN102390370B (en) | Stereoscopic vision based emergency treatment device and method for running vehicles | |
| CN106297330B (en) | Reduce the method and system that pedestrian's street crossing influences plane perceptual signal control efficiency | |
| CN104778274B (en) | A wide range of city road network hourage method of estimation based on sparse GPS data from taxi | |
| CN108417087A (en) | A kind of vehicle safety traffic system and method | |
| CN112740295B (en) | Method and device for detecting complexity of vehicle driving scene | |
| CN103308056A (en) | Road marking detection method | |
| CN105307309B (en) | A kind of illumination control method and system | |
| CN102013159A (en) | High-definition video detection data-based region dynamic origin and destination (OD) matrix acquiring method | |
| CN103177585B (en) | Road turning average travel speed calculating method based on floating car data | |
| CN111027430A (en) | Traffic scene complexity calculation method for intelligent evaluation of unmanned vehicles | |
| CN104408914A (en) | Signal intersection single vehicle stopping delay time estimating method and system based on GPS data | |
| CN110299010A (en) | A kind of information processing method towards bus or train route collaboration roadside device | |
| CN115985104A (en) | Traffic flow prediction device, prediction method and prediction model construction method | |
| CN115165387A (en) | Control method, device and system for testing performance of automatic driving whole vehicle | |
| Wei et al. | Street object detection/tracking for AI city traffic analysis | |
| CN103903432A (en) | Equipment for determining road link congestion state and method | |
| CN116597690B (en) | Highway test scene generation method, equipment and medium for intelligent network-connected automobile | |
| CN104318761B (en) | Highway-scene-based detection and vehicle detection tracking optimization method | |
| Shan et al. | Vehicle collision risk estimation based on RGB-D camera for urban road | |
| TWI712012B (en) | Artificial intelligence traffic detection system | |
| CN111353481A (en) | Road obstacle identification method based on laser point cloud and video image | |
| CN109993081A (en) | A kind of statistical method of traffic flow based on road video and car plate detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information |
Inventor after: He Jin Inventor after: Chang Pengfei Inventor before: Xin Le Inventor before: Gao Jiangjie Inventor before: Fang Shengchao Inventor before: Chen Yangzhou |
|
| CB03 | Change of inventor or designer information | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20170602 Address after: High tech Zone Shanxi city Taiyuan province Jinyang street 030006 No. 161 11 floor Patentee after: Shanxi Heng Road Traffic Survey and Design Consulting Co., Ltd. Address before: 100124 Chaoyang District, Beijing Ping Park, No. 100 Patentee before: Beijing University of Technology |
|
| TR01 | Transfer of patent right | ||
| CP03 | Change of name, title or address |
Address after: 030006 4th floor, building 1, No. 3, Jiahua street, Taiyuan Xuefu Park, comprehensive reform demonstration zone, Taiyuan City, Shanxi Province Patentee after: Zhongke Luheng Engineering Design Co.,Ltd. Address before: 030006 floor 11, No. 161, Jinyang street, high tech Zone, Taiyuan City, Shanxi Province Patentee before: SHANXI LUHENG COMMUNICATIONS SURVEY & DESIGN Co.,Ltd. |
|
| CP03 | Change of name, title or address |