CN103971380A - Pedestrian trailing detection method based on RGB-D - Google Patents
Pedestrian trailing detection method based on RGB-D Download PDFInfo
- Publication number
- CN103971380A CN103971380A CN201410186168.6A CN201410186168A CN103971380A CN 103971380 A CN103971380 A CN 103971380A CN 201410186168 A CN201410186168 A CN 201410186168A CN 103971380 A CN103971380 A CN 103971380A
- Authority
- CN
- China
- Prior art keywords
- image
- depth
- depth image
- rgb
- foreground
- 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
Abstract
Provided is a pedestrian trailing detection method based on RGB-D. The method comprises the steps that a foreground target is detected firstly through a depth image in a monitor area, then a human head target is precisely detected and located according to hair color information and depth outline information of the head and the shoulders, finally, the human head target is trailed through a camshaft algorithm, and then whether a trailing phenomenon occurs or not is judged. According to the method, pedestrians and luggage can be effectively recognized, the pedestrians can be precisely counted, the pedestrians cannot be mistaken for the luggage, the luggage cannot be mistaken for the pedestrians, high detection precision is achieved on the multi-people trailing phenomenon, and the method is especially suitable for the occasions, such as entrance guards, station security checkpoints and enterprise access channels.
Description
Technical field
The invention belongs to intelligent monitoring technology field, particularly relate to a kind of pedestrian based on RGB-D and trail detection method.
Background technology
Along with the fast development of computer vision technique, intelligent monitoring has a wide range of applications in fields such as security protection, safety check and traffic administrations, such as number-plate number inspection, pedestrian or the vehicular traffic field such as detection and visual gate inhibition in violation of rules and regulations.Intellectualized monitoring not only can reduce the cost of manpower monitoring, but also is an important embodiment of urban civilization.
We often saw the report that crushes the accidents such as passenger such as subway gate in the past, trace it to its cause is mainly that existing system cannot be distinguished pedestrian and luggage, and cannot distinguish after swiping the card and have several people to pass through, if passenger swipes the card, pusher luggage and is passed through gate, at this moment gate probably crushed, because can be treated as pedestrian by luggage by mistake.The security check passage such as be all provided with fingerprint in the porch of some companies in addition or check card, but often can there is the people situation that many people trail of checking card.In addition, present stage is many based on two dimensional image for the intelligent monitor system of human detection, is therefore difficult to solve the interference such as the close and shade of illumination variation, color of object.
In No. 201210240640.0th, Chinese invention patent application, a kind of pedestrian detection method that only uses depth information is disclosed, the method is with the difference of depth threshold, adult, child and luggage to be distinguished, and shortcoming is to comparatively difficulty of heavy luggage and child's differentiation.In No. 201110026465.0th, Chinese invention patent application, a kind of pedestrian detection method based on depth information is disclosed, the method is to build disaggregated model by a large amount of depth images are extracted to feature, then the feature of extracting from depth image is input in disaggregated model, whether to comprise human body in judging area, to realize human detection.But it is high that the shortcoming of the method is computation complexity, and cannot carry out the tracking detection of human body target.
Summary of the invention
In order to address the above problem, the object of the present invention is to provide one not only can effectively identify pedestrian and luggage, pedestrian is carried out to accurate metering, and the phenomenon that many people are trailed has compared with the pedestrian based on RGB-D of high measurement accuracy and trails detection method.
In order to achieve the above object, the pedestrian based on RGB-D provided by the invention trails detection method and comprises the following step carrying out in order:
1) set up colour imagery shot and degree of depth camera are arranged side by side directly over monitor channel in monitoring visual angle mode vertically downward, the perpendicular monitoring scene of the straight line at colour imagery shot and degree of depth camera place and monitor channel length direction simultaneously, and utilize RGB image and the depth image in colour imagery shot and degree of depth camera Real-time Collection monitor channel;
2) the background subtraction method of employing based on depth image carried out the foreground extraction of depth image, obtains foreground depth image and prospect RGB image;
3) utilize above-mentioned foreground depth image and prospect RGB image to detect number of people target, determine head zone;
4) in above-mentioned head zone, carry out number of people target following, judge whether thus to trail phenomenon;
In step 2) in, the background subtraction method of described employing based on depth image carried out the foreground extraction of depth image, and the method that obtains foreground depth image and prospect RGB image is:
If background depth image is B
d(x, y), the depth image of present frame is I
d(x, y), RGB image are I
c(x, y), only needs that directly the two is done to difference and can try to achieve foreground mask image M (x, y) and be:
F
d(x,y)=I
d(x,y)·M(x,y)
F
c(x,y)=I
c(x,y)·M(x,y)
In formula, M
tfor decision threshold, F (x, y) is foreground image.F
d(x, y) is foreground depth image, F
c(x, y) is prospect RGB image.
In step 3) in, described utilize foreground depth image and prospect RGB image detect number of people target, the method of determining head zone is: first according to the cluster feature of hair color, from above-mentioned prospect RGB image, extract candidate's head zone, and candidate's head zone is mapped in foreground depth image, then the foreground depth image of candidate's head zone is carried out to rim detection, and use hough change detection class circle ring area wherein, finally remove pseudo-head zone according to priori, obtain contouring head figure.
In step 4) in, described in head zone, carry out number of people target following, judge whether that thus the method for trailing phenomenon is: first using the boundary rectangle of the above-mentioned head zone of obtaining as initial ranging frame, depth value in rectangle frame is added up, obtained the degree of depth histogram model of head target; Then by the degree of depth Histogram backprojection of target in the depth image of institute's tracking frame, obtain the degree of depth probability distribution graph of head target, in this probability distribution graph, the pixel that probability is large shows that this pixel is that the possibility of target is larger; Last according to the tracking results to head target, and judge whether to trail phenomenon in conjunction with application scenarios.
Pedestrian based on RGB-D provided by the invention trails detection method and first detects foreground target by the depth image of guarded region, then according to the depth profile information of hair color information and head, number of people target is accurately detected and located, finally adopt camshift algorithm to follow the tracks of head part's target, and then determine whether and occur trailing phenomenon.This method not only can effectively identify pedestrian and luggage, pedestrian is carried out to accurate metering, can be not by mistake by pedestrian as luggage or by luggage as pedestrian, and the phenomenon that many people are trailed has higher accuracy of detection, is specially adapted to the places such as gate inhibition, security check, station, company's Vomitory.
Brief description of the drawings
Fig. 1 trails the monitoring scene vertical view adopting in detection method for the pedestrian based on RGB-D provided by the invention.
Fig. 2 pedestrian based on RGB-D provided by the invention trails people's head inspecting method process flow diagram in detection method.
Fig. 3 is that the pedestrian based on RGB-D provided by the invention trails in detection method the number of people and follows the tracks of and trail method flow diagram.
Embodiment
The pedestrian based on RGB-D who the utility model is provided below in conjunction with the drawings and specific embodiments trails detection method and is elaborated.
As shown in Fig. 1-Fig. 3, the pedestrian based on RGB-D provided by the invention trails detection method and comprises the following step carrying out in order:
1) set up colour imagery shot 1 and degree of depth camera 2 are arranged side by side directly over monitor channel 3 in monitoring visual angle mode vertically downward as shown in Figure 1, colour imagery shot 1 should be close as much as possible with the distance of degree of depth camera, with the interference that reduces to cause because of different angulars field of view; The perpendicular monitoring scene of the straight line at colour imagery shot 1 and degree of depth camera 2 places and monitor channel 3 length directions simultaneously, and utilize RGB image and the depth image in colour imagery shot 1 and degree of depth camera 2 Real-time Collection monitor channels 3;
Original monitor video only can provide rgb video image, but in image, but lose spatial information, we can only go to distinguish different targets by some feature from RGB color space, however in space cutting apart of target be that connectivity with its space is cut apart.The present invention completes the precise monitoring to pedestrian target in conjunction with RGB colouring information and spatial depth information.The impact of the disturbing factors such as target and background color are close because the obtaining of depth image can not be subject to, illumination variation and shade, so can extract foreground target with degree of precision.Because of eclipse phenomena inevitable but have a strong impact on monitoring result in monitoring field, in order to reduce as far as possible the eclipse phenomena between moving target, this method adopts method shown in Fig. 1, and colour imagery shot 1 and degree of depth camera 2 are arranged on to passage top and monitor with visual angle vertically downward.
2) the background subtraction method of employing based on depth image carried out the foreground extraction of depth image, obtains foreground depth image and prospect RGB image;
Background subtraction method is one of conventional means of foreground detection, and it,, by conventional images relatively and known background image, detects the place that difference is larger and be defined as prospect.Mostly current method is the video flowing based on RGB, and in the time being subject to illumination, shade and affecting in the time that background colour is close with foreground target color, can the serious accuracy that reduces foreground detection.The present invention adopts the background subtraction method based on depth image.Because depth image reaction is the spatial information of scene, the each pixel value in image represents that corresponding object arrives the distance of camera plane.Depth image can not be subject to the impact of shade in scene, illumination and change color.Concrete grammar is as follows:
If background depth image is B
d(x, y), the depth image of present frame is I
d(x, y), RGB image are I
c(x, y), only needs that directly the two is done to difference and can try to achieve foreground mask image M (x, y) and be:
F
d(x,y)=I
d(x,y)·M(x,y)
F
c(x,y)=I
c(x,y)·M(x,y)
In formula, M
tfor decision threshold, F (x, y) is foreground image.F
d(x, y) is foreground depth image, F
c(x, y) is prospect RGB image.
3) utilize above-mentioned foreground depth image and prospect RGB image to detect number of people target, determine head zone;
In the situation that camera vertically sets up, often adopt the method based on hair color and contouring head for the detection of the number of people.Because Asian's hair color is darker, according to statistics, in RGB color space, the R component of hair color is gathered between 0-35.Generally can pass through this color cluster feature extraction head candidate region of hair.In addition, the number of people contour approximation in monitored picture is circular, and a lot of scholars are by judging the exact position of head to the detection of annulus.But traditional contour extraction method is to carry out according to the gradient of gradation of image.In the time that hair color is close with background (such as clothes, monitoring scene) color, be just difficult to extract exactly the profile of the number of people.
The physical segmentation of considering target is to be based upon above the uncontinuity of its degree of depth, and the highest point of the number of people in human body, head shoulder can not be blocked substantially, so its contour feature is comparatively outstanding, from depth information, obviously distinguish over other body parts, therefore can from image, distinguish exactly number of people target.
Therefore the present invention proposes a kind of number of people profile testing method based on depth image, first according to the cluster feature of hair color, from above-mentioned prospect RGB image, extract candidate's head zone, and candidate's head zone is mapped in foreground depth image, then the foreground depth image of candidate's head zone is carried out to rim detection, and use hough change detection class circle ring area wherein, and finally remove pseudo-head zone according to priori, obtain contouring head figure.Concrete grammar is as follows:
According to hair color cluster feature, set the threshold range R ∈ [0,35] of hair R component, from prospect RGB image, extract candidate's head zone.
In formula, F
r(x, y) is prospect RGB image F
cthe R channel components of (x, y), H
m(x, y) is candidate's head zone mask.
1, ask for the boundary rectangle of candidate's head zone mask, and by this rectangle centered by diagonal line intersection point, length and width are expanded respectively to 3 pixel wide, to ensure that rectangle frame can cover the actual physics profile of head completely.Then this rectangle frame is mapped to foreground depth image F
din (x, y), obtain the depth image of candidate's head zone, be made as H here
d(x, y).
2, adopt the depth image H of canny operator to candidate's head zone
d(x, y) carries out rim detection, by suitable detection threshold is set, can obtain the edge contour image of refinement.
3, use the annulus in hough change detection edge contour image.
4, remove pseudo-head zone according to priori, obtain head zone.
Because contouring head is not standard circular profile, but the class circle contour of sub-circular, same head may have multiple output after hough conversion, according to prioris such as camera height and number of people sizes, in the time using hough conversion to carry out annulus detection, the least radius of first setting circle is R
min, maximum radius is R
maxand the center of circle distance R of adjacent annulus
l>R
max.In addition, the present invention extracts the depth value of head zone from multi-amplitude deepness image, asks for its average M.Statistics shows: in the depth image of whole head zone, except subregion is because detecting the inaccurate 0 value region occurring, the depth value of remainder is basicly stable in the scope of M ± 5cm.Obtaining after circle ring area by hough, can pass through annular radii R
i(i is annulus sequence number) tries to achieve its area S
i, then ask for maximum depth value D in circle ring area, and from circle ring area, extract mask with [D-10cm, D] for scope, statistical mask region area S
mi, work as q=S
mi/ S
i>q
0time (q
0for setting threshold), can this annulus of final decision be head zone.
4) in above-mentioned head zone, carry out number of people target following, judge whether thus to trail phenomenon;
Aspect target following, camshift algorithm is really a kind of method of real-time high-efficiency, and it utilizes the colouring information in region, completes the tracking to moving target by the mode of cluster.Traditional camshift algorithm must be by image from RGB color space conversion to hsv color space, and then utilizes the histogram of H component to set up the color probability model of target.But in the time that moving target and background color approach, its testing result is difficult to satisfactory.
The present invention is inspired from the conventional camshift algorithm in motion target tracking field, because in the situation that camera vertically sets up, when pedestrian passes by monitored space, head part is almost constant apart from the distance of camera plane, and the depth value scope that embodiment is people's head region on depth image is almost constant.Therefore the present invention proposes a kind of camshift target tracking method based on depth image, and the H component using in traditional camshift algorithm is replaced with depth component by the method, and then utilize camshift algorithm to carry out target following.First using the boundary rectangle of the above-mentioned head zone of obtaining as initial ranging frame, depth value in rectangle frame is added up, obtained the degree of depth histogram model of head target; Then by the degree of depth Histogram backprojection of target in the depth image of institute's tracking frame, obtain the degree of depth probability distribution graph of head target, in this probability distribution graph, the pixel that probability is large shows that this pixel is that the possibility of target is larger; Last according to the tracking results to head target, and judge whether to trail phenomenon in conjunction with application scenarios.
Concrete grammar is as follows:
1, the minimum boundary rectangle of getting head zone is as search window, if there are multiple head zone respectively they to be carried out to mark.Ask for the depth information histogram of depth image in search box;
2, by Histogram backprojection in former depth image, obtain degree of depth probability distribution graph;
Whether 3, application camshift algorithm is found the region close with search window and is target area in degree of depth probability distribution graph, determine thus and occur trailing.
Claims (4)
1. the pedestrian based on RGB-D trails a detection method, it is characterized in that: it comprises the following step carrying out in order:
1) set up colour imagery shot (1) and degree of depth camera (2) are arranged side by side directly over monitor channel (3) to monitor visual angle mode vertically downward, the perpendicular monitoring scene of the straight line at colour imagery shot (1) and degree of depth camera (2) place and monitor channel (3) length direction simultaneously, and utilize RGB image and the depth image in colour imagery shot (1) and degree of depth camera (2) Real-time Collection monitor channel (3);
2) the background subtraction method of employing based on depth image carried out the foreground extraction of depth image, obtains foreground depth image and prospect RGB image;
3) utilize above-mentioned foreground depth image and prospect RGB image to detect number of people target, determine head zone;
4) in above-mentioned head zone, carry out number of people target following, judge whether thus to trail phenomenon.
2. the pedestrian based on RGB-D according to claim 1 trails detection method, it is characterized in that: in step 2) in, the background subtraction method of described employing based on depth image carried out the foreground extraction of depth image, and the method that obtains foreground depth image and prospect RGB image is:
If background depth image is B
d(x, y), the depth image of present frame is I
d(x, y), RGB image are I
c(x, y), only needs that directly the two is done to difference and can try to achieve foreground mask image M (x, y) and be:
F
d(x,y)=I
d(x,y) M(x,y)
F
c(x,y)=I
c(x,y) M(x,y)
In formula, M
tfor decision threshold, F (x, y) is foreground image.F
d(x, y) is foreground depth image, F
c(x, y) is prospect RGB image.
3. the pedestrian based on RGB-D according to claim 1 trails detection method, it is characterized in that: in step 3) in, described utilize foreground depth image and prospect RGB image detect number of people target, the method of determining head zone is: first according to the cluster feature of hair color, from above-mentioned prospect RGB image, extract candidate's head zone, and candidate's head zone is mapped in foreground depth image, then the foreground depth image of candidate's head zone is carried out to rim detection, and use hough change detection class circle ring area wherein, finally remove pseudo-head zone according to priori, obtain contouring head figure.
4. the pedestrian based on RGB-D according to claim 1 trails detection method, it is characterized in that: in step 4) in, described in head zone, carry out number of people target following, judge whether that thus the method for trailing phenomenon is: first using the boundary rectangle of the above-mentioned head zone of obtaining as initial ranging frame, depth value in rectangle frame is added up, obtained the degree of depth histogram model of head target; Then by the degree of depth Histogram backprojection of target in the depth image of institute's tracking frame, obtain the degree of depth probability distribution graph of head target, in this probability distribution graph, the pixel that probability is large shows that this pixel is that the possibility of target is larger; Last according to the tracking results to head target, and judge whether to trail phenomenon in conjunction with application scenarios.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410186168.6A CN103971380B (en) | 2014-05-05 | 2014-05-05 | Pedestrian based on RGB-D trails detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410186168.6A CN103971380B (en) | 2014-05-05 | 2014-05-05 | Pedestrian based on RGB-D trails detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103971380A true CN103971380A (en) | 2014-08-06 |
CN103971380B CN103971380B (en) | 2016-09-28 |
Family
ID=51240829
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410186168.6A Expired - Fee Related CN103971380B (en) | 2014-05-05 | 2014-05-05 | Pedestrian based on RGB-D trails detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103971380B (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104657993A (en) * | 2015-02-12 | 2015-05-27 | 北京格灵深瞳信息技术有限公司 | Lens shielding detection method and device |
CN104751491A (en) * | 2015-04-10 | 2015-07-01 | 中国科学院宁波材料技术与工程研究所 | Method and device for tracking crowds and counting pedestrian flow |
CN105225230A (en) * | 2015-09-11 | 2016-01-06 | 浙江宇视科技有限公司 | A kind of method and device identifying foreground target object |
CN105741271A (en) * | 2016-01-25 | 2016-07-06 | 上海物联网有限公司 | Method for detecting object in depth image |
CN106096512A (en) * | 2016-05-31 | 2016-11-09 | 上海美迪索科电子科技有限公司 | Utilize the detection device and method that vehicles or pedestrians are identified by depth camera |
CN106384353A (en) * | 2016-09-12 | 2017-02-08 | 佛山市南海区广工大数控装备协同创新研究院 | Target positioning method based on RGBD |
CN106454229A (en) * | 2016-09-27 | 2017-02-22 | 成都理想境界科技有限公司 | Monitoring method, camera device, image processing device and monitoring system |
CN106778655A (en) * | 2016-12-27 | 2017-05-31 | 华侨大学 | A kind of entrance based on human skeleton is trailed and enters detection method |
CN107221058A (en) * | 2017-05-25 | 2017-09-29 | 刘萍 | Intelligent channel barrier system |
CN107230226A (en) * | 2017-05-15 | 2017-10-03 | 深圳奥比中光科技有限公司 | Determination methods, device and the storage device of human body incidence relation |
CN107491712A (en) * | 2016-06-09 | 2017-12-19 | 北京雷动云合智能技术有限公司 | A kind of human body recognition method based on RGB D images |
CN108335308A (en) * | 2017-01-20 | 2018-07-27 | 深圳市祈飞科技有限公司 | A kind of orange automatic testing method, system and intelligent robot retail terminal |
CN108564063A (en) * | 2018-04-27 | 2018-09-21 | 北京华捷艾米科技有限公司 | Centre of the palm localization method based on depth information and system |
JP2018535457A (en) * | 2016-10-25 | 2018-11-29 | シェンチェン ユニバーシティー | Statistical method and apparatus for passersby based on identification of human head top |
CN109688452A (en) * | 2018-12-04 | 2019-04-26 | 深圳市子瑜杰恩科技有限公司 | Pagination Display stage property stacking method and Related product |
CN109977109A (en) * | 2019-04-03 | 2019-07-05 | 深圳市甲易科技有限公司 | A kind of track data cleaning method and adjoint analysis method |
CN110516602A (en) * | 2019-08-28 | 2019-11-29 | 杭州律橙电子科技有限公司 | A kind of public traffice passenger flow statistical method based on monocular camera and depth learning technology |
CN110969747A (en) * | 2019-12-11 | 2020-04-07 | 盛视科技股份有限公司 | Anti-following access control system and anti-following method |
CN111144231A (en) * | 2019-12-09 | 2020-05-12 | 深圳市鸿逸达科技有限公司 | Self-service channel anti-trailing detection method and system based on depth image |
CN111192391A (en) * | 2018-10-25 | 2020-05-22 | 杭州海康威视数字技术股份有限公司 | Pedestrian passageway gate control method and device based on images and/or videos |
CN111723770A (en) * | 2020-06-30 | 2020-09-29 | 四川兴事发门窗有限责任公司 | Anti-trailing gate system and method based on image recognition |
CN113065397A (en) * | 2021-03-02 | 2021-07-02 | 南京苏宁软件技术有限公司 | Pedestrian detection method and device |
WO2022151507A1 (en) * | 2021-01-18 | 2022-07-21 | 深圳市大疆创新科技有限公司 | Movable platform and method and apparatus for controlling same, and machine-readable storage medium |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103514429B (en) * | 2012-06-21 | 2018-06-22 | 夏普株式会社 | Detect the method and image processing equipment of the privileged site of object |
CN102737235B (en) * | 2012-06-28 | 2014-05-07 | 中国科学院自动化研究所 | Head posture estimation method based on depth information and color image |
CN102999892B (en) * | 2012-12-03 | 2015-08-12 | 东华大学 | Based on the depth image of region mask and the intelligent method for fusing of RGB image |
CN103150559B (en) * | 2013-03-01 | 2016-07-06 | 南京理工大学 | Head recognition and tracking method based on Kinect three-dimensional depth image |
-
2014
- 2014-05-05 CN CN201410186168.6A patent/CN103971380B/en not_active Expired - Fee Related
Cited By (32)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104657993A (en) * | 2015-02-12 | 2015-05-27 | 北京格灵深瞳信息技术有限公司 | Lens shielding detection method and device |
CN104751491A (en) * | 2015-04-10 | 2015-07-01 | 中国科学院宁波材料技术与工程研究所 | Method and device for tracking crowds and counting pedestrian flow |
CN104751491B (en) * | 2015-04-10 | 2018-01-23 | 中国科学院宁波材料技术与工程研究所 | A kind of crowd's tracking and people flow rate statistical method and device |
CN105225230A (en) * | 2015-09-11 | 2016-01-06 | 浙江宇视科技有限公司 | A kind of method and device identifying foreground target object |
CN105225230B (en) * | 2015-09-11 | 2018-07-13 | 浙江宇视科技有限公司 | A kind of method and device of identification foreground target object |
CN105741271A (en) * | 2016-01-25 | 2016-07-06 | 上海物联网有限公司 | Method for detecting object in depth image |
CN105741271B (en) * | 2016-01-25 | 2021-11-16 | 上海物联网有限公司 | Method for detecting object in depth image |
CN106096512A (en) * | 2016-05-31 | 2016-11-09 | 上海美迪索科电子科技有限公司 | Utilize the detection device and method that vehicles or pedestrians are identified by depth camera |
CN106096512B (en) * | 2016-05-31 | 2020-08-25 | 上海美迪索科电子科技有限公司 | Detection device and method for recognizing vehicle or pedestrian by using depth camera |
CN107491712A (en) * | 2016-06-09 | 2017-12-19 | 北京雷动云合智能技术有限公司 | A kind of human body recognition method based on RGB D images |
CN106384353A (en) * | 2016-09-12 | 2017-02-08 | 佛山市南海区广工大数控装备协同创新研究院 | Target positioning method based on RGBD |
CN106454229A (en) * | 2016-09-27 | 2017-02-22 | 成都理想境界科技有限公司 | Monitoring method, camera device, image processing device and monitoring system |
JP2018535457A (en) * | 2016-10-25 | 2018-11-29 | シェンチェン ユニバーシティー | Statistical method and apparatus for passersby based on identification of human head top |
CN106778655A (en) * | 2016-12-27 | 2017-05-31 | 华侨大学 | A kind of entrance based on human skeleton is trailed and enters detection method |
CN106778655B (en) * | 2016-12-27 | 2020-05-05 | 华侨大学 | Human body skeleton-based entrance trailing entry detection method |
CN108335308A (en) * | 2017-01-20 | 2018-07-27 | 深圳市祈飞科技有限公司 | A kind of orange automatic testing method, system and intelligent robot retail terminal |
CN107230226A (en) * | 2017-05-15 | 2017-10-03 | 深圳奥比中光科技有限公司 | Determination methods, device and the storage device of human body incidence relation |
CN107221058A (en) * | 2017-05-25 | 2017-09-29 | 刘萍 | Intelligent channel barrier system |
CN108564063A (en) * | 2018-04-27 | 2018-09-21 | 北京华捷艾米科技有限公司 | Centre of the palm localization method based on depth information and system |
CN111192391B (en) * | 2018-10-25 | 2022-09-23 | 杭州海康威视数字技术股份有限公司 | Pedestrian passageway gate control method and device based on images and/or videos |
CN111192391A (en) * | 2018-10-25 | 2020-05-22 | 杭州海康威视数字技术股份有限公司 | Pedestrian passageway gate control method and device based on images and/or videos |
CN109688452A (en) * | 2018-12-04 | 2019-04-26 | 深圳市子瑜杰恩科技有限公司 | Pagination Display stage property stacking method and Related product |
CN109977109A (en) * | 2019-04-03 | 2019-07-05 | 深圳市甲易科技有限公司 | A kind of track data cleaning method and adjoint analysis method |
CN110516602A (en) * | 2019-08-28 | 2019-11-29 | 杭州律橙电子科技有限公司 | A kind of public traffice passenger flow statistical method based on monocular camera and depth learning technology |
CN111144231A (en) * | 2019-12-09 | 2020-05-12 | 深圳市鸿逸达科技有限公司 | Self-service channel anti-trailing detection method and system based on depth image |
WO2021114765A1 (en) * | 2019-12-09 | 2021-06-17 | 深圳市鸿逸达科技有限公司 | Depth image-based method and system for anti-trailing detection of self-service channel |
CN110969747A (en) * | 2019-12-11 | 2020-04-07 | 盛视科技股份有限公司 | Anti-following access control system and anti-following method |
CN111723770A (en) * | 2020-06-30 | 2020-09-29 | 四川兴事发门窗有限责任公司 | Anti-trailing gate system and method based on image recognition |
CN111723770B (en) * | 2020-06-30 | 2020-12-18 | 四川兴事发门窗有限责任公司 | Anti-trailing gate system and method based on image recognition |
WO2022151507A1 (en) * | 2021-01-18 | 2022-07-21 | 深圳市大疆创新科技有限公司 | Movable platform and method and apparatus for controlling same, and machine-readable storage medium |
CN113065397A (en) * | 2021-03-02 | 2021-07-02 | 南京苏宁软件技术有限公司 | Pedestrian detection method and device |
CN113065397B (en) * | 2021-03-02 | 2022-12-23 | 南京苏宁软件技术有限公司 | Pedestrian detection method and device |
Also Published As
Publication number | Publication date |
---|---|
CN103971380B (en) | 2016-09-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103971380A (en) | Pedestrian trailing detection method based on RGB-D | |
Luvizon et al. | A video-based system for vehicle speed measurement in urban roadways | |
CN101794385B (en) | Multi-angle multi-target fast human face tracking method used in video sequence | |
CN105718870B (en) | Based on the preceding roadmarking extracting method to camera in automatic Pilot | |
CN102542289B (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
CN103279736B (en) | A kind of detection method of license plate based on multi-information neighborhood ballot | |
CN102880863B (en) | Method for positioning license number and face of driver on basis of deformable part model | |
CN103150549B (en) | A kind of road tunnel fire detection method based on the early stage motion feature of smog | |
CN104966049B (en) | Lorry detection method based on image | |
CN106446926A (en) | Transformer station worker helmet wear detection method based on video analysis | |
CN105488454A (en) | Monocular vision based front vehicle detection and ranging method | |
CN104778444A (en) | Method for analyzing apparent characteristic of vehicle image in road scene | |
CN104134079A (en) | Vehicle license plate recognition method based on extremal regions and extreme learning machine | |
CN103310194A (en) | Method for detecting head and shoulders of pedestrian in video based on overhead pixel gradient direction | |
CN104573697B (en) | Building hoist car demographic method based on Multi-information acquisition | |
CN103136528A (en) | Double-edge detection based vehicle license plate identification method | |
CN102855508B (en) | Opening type campus anti-following system | |
CN102999749A (en) | Intelligent safety belt regulation violation event detecting method based on face detection | |
CN104183142A (en) | Traffic flow statistics method based on image visual processing technology | |
CN103034852A (en) | Specific color pedestrian detecting method in static video camera scene | |
CN102915433A (en) | Character combination-based license plate positioning and identifying method | |
CN102831420A (en) | Circular traffic sign positioning method based on color information and randomized circle detection | |
CN105184301B (en) | A kind of method that vehicle heading is differentiated using four-axle aircraft | |
CN104778727A (en) | Floating car counting method based on video monitoring processing technology | |
CN103390151A (en) | Face detection method and device |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160928 Termination date: 20190505 |
|
CF01 | Termination of patent right due to non-payment of annual fee |