CN104156983A - Public transport passenger flow statistical method based on video image processing - Google Patents
Public transport passenger flow statistical method based on video image processing Download PDFInfo
- Publication number
- CN104156983A CN104156983A CN201410380811.9A CN201410380811A CN104156983A CN 104156983 A CN104156983 A CN 104156983A CN 201410380811 A CN201410380811 A CN 201410380811A CN 104156983 A CN104156983 A CN 104156983A
- Authority
- CN
- China
- Prior art keywords
- passenger
- frame
- target
- video image
- surf
- 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
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a public transport passenger flow statistical method based on video image processing. The public transport passenger flow statistical method comprises the following steps: pretreatment is performed on each selected frame of video image to obtain outline of each foreground in the each frame of video image; some video images are utilized to train a passenger target detection classifier so as to obtain a strong classifier; when passenger flow statistics is carried out, the trained strong classifier is utilized to traverse a current-frame video image window subjected to foreground extraction on locational space with different sizes, and if the window is distinguished as a passenger target, the position and the current dimension are recorded; otherwise, the position and the current dimension are discarded; a passenger target chain is established for storing all passenger targets; the SURF algorithm is adopted to perform SURF feature extraction and feature point matching on every two adjacent frame video images; the passenger target chain is updated according to different conditions so as to track the passenger targets. According to the invention, on the conditions that the clothes colors of passengers are similar to the color of a background, objects in the background are similar to the outlines of passengers, and the weather or the illumination changes, the public transport passenger flow statistics can still be carried out.
Description
Affiliated technical field
The present invention relates to a kind of public traffice passenger flow statistical method.
Background technology
Along with China's population constantly increases, urban-rural integration process constantly advances, and the traffic of China is faced with stern challenge.Greatly developing public transport is the effective means of solving urban traffic blocking, is conducive to rationally utilize urban transportation resource.Bus passenger flow statistics can reflect real bus operation situation, is the basis of the bus dispatching scheme of formulation science, is to provide the guarantee of high-quality public transport service.
The method of bus passenger flow statistics has a lot, and wherein relatively main stream approach has infrared electronic technology statistic law both at home and abroad,, at upper and lower car door place installation infrared sniffer, utilizes the infrared ray of human body radiation to carry out the detection of passenger flow counting and passenger behavior direction.There are some drawbacks in the mode of this passenger flow statistics, as large in climate impact, and when, passenger serious shielding crowded personnel, statistical error is large etc.
Use based on the method for video image processing be camera collection that bus is arranged on to roof vertical direction to video process, thereby the method for statistics bus passenger flow quantity.The method has effectively reduced passenger and has caused the impact of blocking for passenger flow statistics because of crowded, and statistical precision is high, is considered to one of method being best suited for bus passenger flow statistics.But bus camera collection to video affected by light larger, this can reduce statistical precision.How reducing light changes the statistical precision that causes and changes and obtained paying close attention to widely.
Summary of the invention
The object of the invention is to overcome the above-mentioned deficiency of prior art, provide one to ensure mutually to block and light changes in obvious situation passenger, still have the public traffice passenger flow statistical method of higher verification and measurement ratio and tracking rate.Statistical method provided by the invention, utilize edge feature and the local invariant feature of human head and shoulder portion, first treat detected image and carry out passenger's Target Segmentation, then will combine based on a human detection algorithm of shoulder HOG feature and the track algorithm based on SURF feature, realize the detection and tracking to the passenger that gets on or off the bus, and according to passenger's direction of motion, realize bus passenger flow counting, can further improve bus passenger flow counting precision, reduce counting cost.
Technical scheme of the present invention is as follows:
Based on a public traffice passenger flow statistical method for video image processing, utilize the camera collection that is arranged on bus roof vertical direction to video process, step is as follows:
1) selected each frame video image is carried out to pre-service, obtain the profile of each prospect in each frame video image;
2) utilize some video images, carry out based on an occupant detection for shoulder HOG feature, training passenger target detection sorter, obtains strong classifier;
3) in the time carrying out passenger flow statistics, if present frame is N frame, utilize the strong classifier obtaining after training on different size positions spaces, to the current frame video image window traversal after foreground extraction, if this window is identified as passenger's target, record position and current yardstick, otherwise abandon; Set up passenger's object chain, for storing all passenger's targets;
4) adopt SURF algorithm every adjacent two frame video images are carried out to SURF feature extraction and Feature Points Matching;
5) the SURF characteristic matching result of consecutive frame, can produce four kinds of situations: passenger does not change, in image without passenger, occur new passenger's target and passenger's target temporarily missing, adopt following concrete judgment criteria to judge, and carry out the renewal of passenger's object chain, thus realize the tracking of passenger's target:
A. passenger does not change: N frame and N+1 frame SURF characteristic matching, and N frame has occupant detection result, upgrade passenger's object chain by the tracking results of N+1 frame;
B. in image without passenger: N frame and N+1 frame SURF characteristic matching, and N frame is without occupant detection result.Do not upgrade passenger's object chain;
C. there is new passenger's target: N frame does not mate with N+1 frame SURF feature, and N frame is without occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame has occupant detection result; With the tracking results renewal passenger object chain of N+2 frame.
D. passenger's target is temporarily missing: N frame does not mate with N+1 frame SURF feature, and N frame has occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame is without occupant detection result.If exceed timing range, think that target leaves, it is rejected from object chain.
Wherein, step 1) in, selected each frame video image is carried out to pretreated method can be as follows: adopt Otsu method self-adaptation to choose background difference threshold value, to and carry out adaptive threshold background subtraction divisional processing; Differentiated video image is carried out to denoising with median filtering method, then image is corroded to operation, then image is carried out to expansive working, complete the morphology processing of differentiated video image; The method that adopts polygon to approach profile is found out the profile of each prospect in difference image.
Step 2) in, carry out based on an occupant detection for shoulder HOG feature, training passenger target detection sorter, the method that obtains strong classifier can be as follows: the positive sample that manually cutting comprises passenger's shoulder and the negative sample that does not comprise passenger's shoulder, and all samples are zoomed to same size; Extract the HOG feature of all positive negative samples, and align negative sample and give label, all positive sample labelings are 1, and all negative samples are labeled as 0; Adopt Adaboost algorithm, train some SVM Weak Classifiers for different sample sets, and these Weak Classifiers are joined together, form a final strong classifier.
The present invention can be similar to background color in human body clothes color, have in the situations such as object, weather or the illumination variation of similar human body contour outline in background, still can accurately carry out bus passenger flow statistics, in reducing passenger's target probability of false detection, also reduce the cost of bus passenger flow statistics.
Brief description of the drawings
Fig. 1 is basic flow sheet of the present invention.
Fig. 2 is passenger's target detection principle schematic of the present invention.
Fig. 3 is passenger's target following principle schematic of the present invention.
Embodiment
Below in conjunction with Fig. 1 to 3 and embodiment, the present invention will be described, and public traffice passenger flow statistical method of the present invention is as follows:
1) be written into and be installed on the video image that the camera collection of bus roof vertical direction arrives, extract 1 frame every 5 frames, as the image to be detected of passenger flow counting.
2) treat detected image and carry out passenger's Target Segmentation, comprise following steps.
A. adopt background subtraction point-score to process image to be detected, every 1 two field picture and background image carry out difference, and according to threshold value, the difference image obtaining are carried out to binaryzation, and then judge that a certain pixel belongs to prospect or background.Binary-state threshold is chosen by Otsu method self-adaptation.
B. difference image is carried out to morphology processing, use median filtering algorithm to difference image denoising.Image is corroded to operation, tighten image-region, the part that is less than structural element is removed, and then do expansive working, the region after expansion deflation, fills up the cavity in object after Image.
C. difference image is carried out to foreground extraction, obtain passenger target area.The method that adopts polygon to approach profile is found out each prospect profile in difference image, then sets area threshold, if contour area is greater than this threshold value, draws the minimum rectangle frame that comprises this profile, otherwise abandons.The minimum rectangle frame obtaining is passenger target area.
3) in passenger target area, carry out based on an occupant detection for shoulder HOG feature, concrete steps are as follows.
A. train passenger's target detection sorter.The positive sample that manually cutting comprises passenger's shoulder and the negative sample that does not comprise passenger's shoulder, and all samples are zoomed to same size; Extract the HOG feature of all positive negative samples, and align negative sample and give label, all positive sample labelings are 1, and all negative samples are labeled as 0; Adopt Adaboost algorithm, train some SVM Weak Classifiers for different sample sets, and these Weak Classifiers are joined together, form a final strong classifier
B. utilize strong classifier to carry out passenger's target detection.On different yardsticks and locational space, treat the foreground area of detected image and carry out window traversal, if this window is identified as passenger's target, records its position and current yardstick, otherwise abandon.
4) extract the SURF feature of consecutive frame, and carry out Feature Points Matching.Judge whether that according to Feature Points Matching situation passenger getting off car appears, whether has in new passenger's target, thereby upgraded passenger's object chain, realized passenger flow counting.Taking N frame, N+1 frame and N+2 frame as example, be elaborated and how carry out the renewal of passenger's object chain.First, using N frame as target area, N+1 frame, as region of search, extracts respectively both SURF unique points, completes after Feature Points Matching, and using N+1 frame as target area, N+2 frame is as region of search, then repeats above-mentioned steps.There will be situation in following 4.
A. N frame and N+1 frame SURF characteristic matching, and N frame has occupant detection result, shows that in image, passenger does not change, and upgrades passenger's object chain by the tracking results of N+1 frame.
B. N frame and N+1 frame SURF characteristic matching, and N frame is without occupant detection result, shows in image, without passenger, not upgrade passenger's object chain.
C. N frame does not mate with N+1 frame SURF feature, and N frame is without occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame has occupant detection result, shows to have new passenger's target to occur, with the tracking results renewal passenger object chain of N+2 frame.
D. N frame does not mate with N+1 frame SURF feature, and N frame has occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame, without occupant detection result, shows to have passenger's target temporarily to disappear, if exceed timing range, thinks that target leaves, and it is rejected from object chain.
Claims (3)
1. the public traffice passenger flow statistical method based on video image processing, utilize the camera collection that is arranged on bus roof vertical direction to video process, step is as follows:
1) selected each frame video image is carried out to pre-service, obtain the profile of each prospect in each frame video image;
2) utilize some video images, carry out based on an occupant detection for shoulder HOG feature, training passenger target detection sorter, obtains strong classifier;
3) in the time carrying out passenger flow statistics, if present frame is N frame, utilize the strong classifier obtaining after training on different size positions spaces, to the current frame video image window traversal after foreground extraction, if this window is identified as passenger's target, record position and current yardstick, otherwise abandon; Set up passenger's object chain, for storing all passenger's targets;
4) adopt SURF algorithm every adjacent two frame video images are carried out to SURF feature extraction and Feature Points Matching;
5) the SURF characteristic matching result of consecutive frame, can produce four kinds of situations: passenger does not change, in image without passenger, occur new passenger's target and passenger's target temporarily missing, adopt following concrete judgment criteria to judge, and carry out the renewal of passenger's object chain, thus realize the tracking of passenger's target:
A. passenger does not change: N frame and N+1 frame SURF characteristic matching, and N frame has occupant detection result, upgrade passenger's object chain by the tracking results of N+1 frame;
B. in image without passenger: N frame and N+1 frame SURF characteristic matching, and N frame is without occupant detection result.Do not upgrade passenger's object chain;
C. there is new passenger's target: N frame does not mate with N+1 frame SURF feature, and N frame is without occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame has occupant detection result; With the tracking results renewal passenger object chain of N+2 frame.
D. passenger's target is temporarily missing: N frame does not mate with N+1 frame SURF feature, and N frame has occupant detection result; N+1 frame and N+2 frame SURF characteristic matching, N+1 frame is without occupant detection result.If exceed timing range, think that target leaves, it is rejected from object chain.
2. passenger flow statistical method according to claim 1, it is characterized in that, in step 1, selected each frame video image is carried out to pretreated method as follows: adopt Otsu method self-adaptation to choose background difference threshold value, to and carry out adaptive threshold background subtraction divisional processing; Differentiated video image is carried out to denoising with median filtering method, then image is corroded to operation, then image is carried out to expansive working, complete the morphology processing of differentiated video image; The method that adopts polygon to approach profile is found out the profile of each prospect in difference image.
3. passenger flow statistical method according to claim 1, it is characterized in that, in step 2, carry out based on an occupant detection for shoulder HOG feature, training passenger target detection sorter, the method that obtains strong classifier is as follows: the positive sample that manually cutting comprises passenger's shoulder and the negative sample that does not comprise passenger's shoulder, and all samples are zoomed to same size; Extract the HOG feature of all positive negative samples, and align negative sample and give label, all positive sample labelings are 1, and all negative samples are labeled as 0; Adopt Adaboost algorithm, train some SVM Weak Classifiers for different sample sets, and these Weak Classifiers are joined together, form a final strong classifier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410380811.9A CN104156983A (en) | 2014-08-05 | 2014-08-05 | Public transport passenger flow statistical method based on video image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410380811.9A CN104156983A (en) | 2014-08-05 | 2014-08-05 | Public transport passenger flow statistical method based on video image processing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104156983A true CN104156983A (en) | 2014-11-19 |
Family
ID=51882472
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410380811.9A Pending CN104156983A (en) | 2014-08-05 | 2014-08-05 | Public transport passenger flow statistical method based on video image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104156983A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104821025A (en) * | 2015-04-29 | 2015-08-05 | 广州运星科技有限公司 | Passenger flow detection method and detection system thereof |
| CN105243420A (en) * | 2015-10-16 | 2016-01-13 | 郑州天迈科技股份有限公司 | Accurate statistical method of bus passenger flow |
| CN106250820A (en) * | 2016-07-20 | 2016-12-21 | 华南理工大学 | A kind of staircase mouth passenger flow based on image procossing is blocked up detection method |
| CN106548163A (en) * | 2016-11-25 | 2017-03-29 | 青岛大学 | Method based on TOF depth camera passenger flow countings |
| CN106570440A (en) * | 2015-10-09 | 2017-04-19 | 株式会社日立制作所 | People counting method and people counting device based on image analysis |
| CN106709906A (en) * | 2016-12-08 | 2017-05-24 | 武汉大学 | Background difference and foreground detection method |
| CN108563998A (en) * | 2018-03-16 | 2018-09-21 | 新智认知数据服务有限公司 | Vivo identification model training method, biopsy method and device |
| CN108764181A (en) * | 2018-05-31 | 2018-11-06 | 北京学之途网络科技有限公司 | A kind of passenger flow statistical method and device, computer readable storage medium |
| CN109344743A (en) * | 2018-09-14 | 2019-02-15 | 广州市浪搏科技有限公司 | A kind of monitor video data processing implementation method |
| CN109919593A (en) * | 2019-03-21 | 2019-06-21 | 京东方科技集团股份有限公司 | Settlement method and system, mobile unit, computer equipment and medium |
| CN110276769A (en) * | 2018-03-13 | 2019-09-24 | 上海狮吼网络科技有限公司 | Live content localization method in a kind of video picture-in-pictures framework |
| CN110765940A (en) * | 2019-10-22 | 2020-02-07 | 杭州姿感科技有限公司 | Target object statistical method and device |
| CN110969131A (en) * | 2019-12-04 | 2020-04-07 | 大连理工大学 | Subway people flow counting method based on scene flow |
| CN111476061A (en) * | 2019-01-23 | 2020-07-31 | 阿里巴巴集团控股有限公司 | Counting processing method, device and system in commodity production and electronic equipment |
| CN112580633A (en) * | 2020-12-25 | 2021-03-30 | 博大视野(厦门)科技有限公司 | Public transport passenger flow statistical device and method |
| CN114724123A (en) * | 2022-03-30 | 2022-07-08 | 东南大学 | Bus passenger flow statistical method based on vehicle-mounted monitoring video |
| CN116432871A (en) * | 2023-06-13 | 2023-07-14 | 北京化工大学 | Bus dispatching optimization method based on AdaBoost algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101794382A (en) * | 2010-03-12 | 2010-08-04 | 华中科技大学 | Method for counting passenger flow of buses in real time |
| CN103310194A (en) * | 2013-06-07 | 2013-09-18 | 太原理工大学 | Method for detecting head and shoulders of pedestrian in video based on overhead pixel gradient direction |
| WO2014094627A1 (en) * | 2012-12-19 | 2014-06-26 | Huawei Technologies Co., Ltd. | System and method for video detection and tracking |
-
2014
- 2014-08-05 CN CN201410380811.9A patent/CN104156983A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101794382A (en) * | 2010-03-12 | 2010-08-04 | 华中科技大学 | Method for counting passenger flow of buses in real time |
| WO2014094627A1 (en) * | 2012-12-19 | 2014-06-26 | Huawei Technologies Co., Ltd. | System and method for video detection and tracking |
| CN103310194A (en) * | 2013-06-07 | 2013-09-18 | 太原理工大学 | Method for detecting head and shoulders of pedestrian in video based on overhead pixel gradient direction |
Non-Patent Citations (3)
| Title |
|---|
| ZHOU DAN等: "A Robust Object Tracking Algorithm Based on SURF", 《2013 INTERNATIONAL CONFERENCE ON WIRELESS COMMUNICATIONS & SIGNAL PROCESSING(WCSP)》 * |
| 林可薇: "以HOG为基础的AdaBoost方法做行人的头部和肩部侦测", 《清华大学博硕士论文全文检索系统》 * |
| 谢迪: "智能视频监控系统中若干检测与跟踪算法的研究", 《中国博士论文全文数据库》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104821025A (en) * | 2015-04-29 | 2015-08-05 | 广州运星科技有限公司 | Passenger flow detection method and detection system thereof |
| CN104821025B (en) * | 2015-04-29 | 2018-01-19 | 广州运星科技有限公司 | Passenger flow detection method and its detecting system |
| CN106570440A (en) * | 2015-10-09 | 2017-04-19 | 株式会社日立制作所 | People counting method and people counting device based on image analysis |
| CN105243420A (en) * | 2015-10-16 | 2016-01-13 | 郑州天迈科技股份有限公司 | Accurate statistical method of bus passenger flow |
| CN105243420B (en) * | 2015-10-16 | 2018-03-20 | 郑州天迈科技股份有限公司 | Bus passenger flow precise statistical method |
| CN106250820A (en) * | 2016-07-20 | 2016-12-21 | 华南理工大学 | A kind of staircase mouth passenger flow based on image procossing is blocked up detection method |
| CN106250820B (en) * | 2016-07-20 | 2019-06-18 | 华南理工大学 | A kind of staircase mouth passenger flow congestion detection method based on image procossing |
| CN106548163B (en) * | 2016-11-25 | 2019-04-16 | 青岛大学 | Method based on TOF depth camera passenger flow counting |
| CN106548163A (en) * | 2016-11-25 | 2017-03-29 | 青岛大学 | Method based on TOF depth camera passenger flow countings |
| CN106709906B (en) * | 2016-12-08 | 2019-07-23 | 武汉大学 | A kind of method of background difference and prospect detection |
| CN106709906A (en) * | 2016-12-08 | 2017-05-24 | 武汉大学 | Background difference and foreground detection method |
| CN110276769A (en) * | 2018-03-13 | 2019-09-24 | 上海狮吼网络科技有限公司 | Live content localization method in a kind of video picture-in-pictures framework |
| CN110276769B (en) * | 2018-03-13 | 2023-07-11 | 上海薇龙文化传播有限公司 | Live broadcast content positioning method in video picture-in-picture architecture |
| CN108563998A (en) * | 2018-03-16 | 2018-09-21 | 新智认知数据服务有限公司 | Vivo identification model training method, biopsy method and device |
| CN108764181A (en) * | 2018-05-31 | 2018-11-06 | 北京学之途网络科技有限公司 | A kind of passenger flow statistical method and device, computer readable storage medium |
| CN108764181B (en) * | 2018-05-31 | 2021-06-08 | 北京学之途网络科技有限公司 | Passenger flow statistical method and device and computer readable storage medium |
| CN109344743A (en) * | 2018-09-14 | 2019-02-15 | 广州市浪搏科技有限公司 | A kind of monitor video data processing implementation method |
| CN109344743B (en) * | 2018-09-14 | 2023-07-25 | 广州市浪搏科技有限公司 | Method for realizing monitoring video data processing |
| CN111476061A (en) * | 2019-01-23 | 2020-07-31 | 阿里巴巴集团控股有限公司 | Counting processing method, device and system in commodity production and electronic equipment |
| CN111476061B (en) * | 2019-01-23 | 2023-05-02 | 阿里巴巴集团控股有限公司 | Counting processing method, device and system in commodity production and electronic equipment |
| US11280629B2 (en) | 2019-03-21 | 2022-03-22 | Boe Technology Group Co., Ltd. | Method for determining trip of user in vehicle, vehicular device, and medium |
| CN109919593A (en) * | 2019-03-21 | 2019-06-21 | 京东方科技集团股份有限公司 | Settlement method and system, mobile unit, computer equipment and medium |
| CN110765940A (en) * | 2019-10-22 | 2020-02-07 | 杭州姿感科技有限公司 | Target object statistical method and device |
| CN110969131A (en) * | 2019-12-04 | 2020-04-07 | 大连理工大学 | Subway people flow counting method based on scene flow |
| CN112580633A (en) * | 2020-12-25 | 2021-03-30 | 博大视野(厦门)科技有限公司 | Public transport passenger flow statistical device and method |
| CN112580633B (en) * | 2020-12-25 | 2024-03-01 | 博大视野(厦门)科技有限公司 | Public transport passenger flow statistics device and method based on deep learning |
| CN114724123A (en) * | 2022-03-30 | 2022-07-08 | 东南大学 | Bus passenger flow statistical method based on vehicle-mounted monitoring video |
| CN114724123B (en) * | 2022-03-30 | 2024-04-23 | 东南大学 | Bus passenger flow statistics method based on vehicle-mounted monitoring video |
| CN116432871A (en) * | 2023-06-13 | 2023-07-14 | 北京化工大学 | Bus dispatching optimization method based on AdaBoost algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104156983A (en) | Public transport passenger flow statistical method based on video image processing | |
| CN106878674B (en) | A kind of parking detection method and device based on monitor video | |
| JP6570731B2 (en) | Method and system for calculating passenger congestion | |
| CN105427626B (en) | A kind of statistical method of traffic flow based on video analysis | |
| CN101872546B (en) | Video-based method for rapidly detecting transit vehicles | |
| CN103077617B (en) | Based on crossing intelligent traffic light supervisory systems and the method for computer vision | |
| CN101382997B (en) | Vehicle detecting and tracking method and device at night | |
| CN102509291B (en) | Pavement disease detecting and recognizing method based on wireless online video sensor | |
| CN104036262B (en) | A kind of method and system of LPR car plates screening identification | |
| CN103971380A (en) | Pedestrian trailing detection method based on RGB-D | |
| CN107808392B (en) | Automatic tracking and positioning method and system for security check vehicle in open scene | |
| CN104298976A (en) | License plate detection method based on convolutional neural network | |
| CN104657724A (en) | Method for detecting pedestrians in traffic videos | |
| CN103268489A (en) | Motor vehicle plate identification method based on sliding window searching | |
| Guo et al. | Image-based seat belt detection | |
| CN104881661B (en) | Vehicle checking method based on structural similarity | |
| CN103049788B (en) | Based on space number for the treatment of object detection system and the method for computer vision | |
| CN104268898A (en) | Method for detecting density of passengers in bus on basis of image analysis | |
| CN107622494A (en) | Towards the vehicle detection at night and tracking of traffic video | |
| CN105260735A (en) | License number character segmentation method based on typical position | |
| CN106951820B (en) | Passenger flow statistical method based on annular template and ellipse fitting | |
| Sheng et al. | Real-time anti-interference location of vehicle license plates using high-definition video | |
| CN103778790B (en) | A kind of vehicle flowrate square wave statistic law based on video sequence | |
| CN113673614B (en) | Metro tunnel foreign matter intrusion detection device and method based on machine vision | |
| CN103295003A (en) | Vehicle detection method based on multi-feature fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141119 |
|
| WD01 | Invention patent application deemed withdrawn after publication |