CN103279791A - Pedestrian counting method based on multiple features - Google Patents
Pedestrian counting method based on multiple features Download PDFInfo
- Publication number
- CN103279791A CN103279791A CN2013102111396A CN201310211139A CN103279791A CN 103279791 A CN103279791 A CN 103279791A CN 2013102111396 A CN2013102111396 A CN 2013102111396A CN 201310211139 A CN201310211139 A CN 201310211139A CN 103279791 A CN103279791 A CN 103279791A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- image
- prospect
- hot spot
- model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000004215 lattice model Methods 0.000 claims abstract description 12
- 230000033001 locomotion Effects 0.000 claims abstract description 9
- 238000001514 detection method Methods 0.000 claims description 19
- 239000000284 extract Substances 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 16
- 230000008878 coupling Effects 0.000 claims description 11
- 238000010168 coupling process Methods 0.000 claims description 11
- 238000005859 coupling reaction Methods 0.000 claims description 11
- 230000003287 optical effect Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000012528 membrane Substances 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000008901 benefit Effects 0.000 abstract description 3
- 230000002146 bilateral effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 15
- 238000006116 polymerization reaction Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000001931 thermography Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000009017 pursuit movement Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a pedestrian counting method based on multiple features. A background lattice model and a capture prospect lattice are set, so the position of the light spot of the background lattice model and the position of the light spot of the capture prospect lattice are compared, a prospect light spot offset model is set, then a light spot offset local maximum value point area in the prospect light spot offset model is extracted to serve as a head area of a pedestrian and is clustered to a head block mass, a motion trail is traced finally, and the number of people is determined and recorded based on motion features. The pedestrian counting method is based on light stream computing, renewal of a pedestrian individual track is achieved by means of intra-frame block mass bilateral matching, and statistics of the number of people is accurate and reliable. The pedestrian counting method is further based on sparse depth data, and improves the execution speed of an algorithm compared with a method based on a dense depth. At last, relative depth acquisition equipment adopted in the pedestrian counting method is low in cost, so the pedestrian counting method has a greater advantage in cost.
Description
Technical field
The present invention relates to video passenger flow statistical technique field, particularly a kind of pedestrian's computing method based on many features.
Background technology
Aspects such as the passenger flow information that the video passenger flow statistical technique provides can be dispatched for Industry Personnel, resource distribution, management tactics, security monitoring provide the data support.Early stage passenger flow statistics can only adopt meetings such as complicate statistics, bill statistics, mechanical means the pedestrian to be caused traditional way of contact of interference.Brought into use electronic trigger or photoelectric sensor afterwards, these equipment are realized comparatively simple, but poor effect under the situations such as, pedestrian's serious shielding big in people's current density.Pedestrian's method of counting based on video can carry out demographics in real time, accurate, glitch-free, becomes lot of domestic and foreign scholar's research focus.Though in the last few years, video pedestrian method of counting is all being obtained significant progress aspect real-time and the accuracy, but still there are a lot of challenges in this technology: (1) since common camera to catch the scene coloured image, foreground detection is subject to the influence of environmental baselines such as illumination variation, shade is serious, scene is complicated and changeable.(2) under or the scene that pedestrian behavior is unusual intensive the pedestrian, be difficult to pedestrian's agglomerate is effectively cut apart or number is estimated, have a strong impact on the accuracy rate of algorithm.(3) some pedestrian detection, track algorithm have proposed requirement for the resolution of video image, but in real world applications, for the consideration to equipment cost, the resolution of video image is generally on the low side.
Present many features pedestrian method of counting mainly contains following both direction:
1) combination of thermal imaging and video image: thermal imaging system characterizes environment temperature on every side with the form of image, and the distance of range sensor is more near, the object that temperature is more high, and the sign in thermal imaging is more obvious.Because the limited coverage area of single-sensor often has only sensor array to play good booster action to video image, such total system cost is tens times even tens times of video pedestrian number system, is not suitable for widespread adoption.
2) combination of scene depth and two dimensional image: because the application of the depth information of scene in pedestrian detection is followed the tracks of has: (1) human body has certain height apart from ground, makes things convenient for the target area to extract; (2) in stereoscopic model, can orient head position fast, be easy to congested cutting apart; (3) be not subjected to advantages such as light variation, shade interference.At the deficiency of existing video pedestrian counting algorithm, many researchs have all proposed the video pedestrian method of counting based on depth information.
The acquiring way of scene depth mainly contains: binocular camera, infrared stereoscopic camera.And the method that combines with camera model of infrared emission model.These methods have different separately suitable scenes, and corresponding cost cost is different.Go out stereoscopic model by combining image pixel grey scale information and pixel depth information architecture, can solve effectively in the scene that light changes, shade disturbs and FAQs such as congested based on stereoscopic model.Therefore, having important use for the research of many features pedestrian counting algorithm of the combining image degree of depth and half-tone information is worth.
Need in actual applications to select proper model and equipment according to statistical precision requirement, scene restriction, cost budgeting etc., following table has provided three kinds of contrasts that the degree of depth is obtained model.
Present a kind of pedestrian's method of counting that merges the degree of depth and video features, this method is made of three modules: depth data recovery, human detection, count tracking.Recover in the module at depth data, the author only uses interpolation algorithm to remedy the black hole, but at first merges the marginal information in depth image and the video image, finishes the compensation of disappearance marginal information.In human detection module, use two layers of tangent plane model as shown below to detect potential pedestrian's head and torso area, determine human region by the effective ratio that detects head and trunk area.
This method is mated according to formula based on depth information and video information at last:
M(a,b)＝αρ(h
a,h
b)+(1-α)ζ(D
a,D
b)
Wherein, ρ (h
a, h
b) be histogram h
aAnd h
bBhattacharyya(Pasteur) distance.ζ (D
a, D
b) be the degree of depth similarity of target a and b.The M value is between 0 to 1, and it is more big to be worth more big expression target matching degree.
Also has pedestrian's method of counting, namely use the infrared transmitter device to add camera model and obtain each regional relative height data of pedestrian's health, the infrared image that camera acquisition arrives, wherein highlighted part is infrared radiation in the zone of human body, and the height of highlighted part is depicted as height accumulated time figure as the relative height of target corresponding region.In depth map, area to be cut apart greater than the agglomerate of certain threshold value, algorithm has calculated the VG (vertical gradient) of pedestrian's agglomerate in the height image, and counts local average gradient curve figure, and the peak value number in the curve map is the pedestrian's number that comprises in the agglomerate.Simultaneously, algorithm is also made judgement based on the height fitting function in the height map agglomerate to the direct of travel of pedestrian's agglomerate.
At present there are following defective in some related art scheme and method: (1) since common camera to catch the scene coloured image, foreground detection is subject to the influence of environmental baselines such as illumination variation, shade is serious, scene is complicated and changeable.(2) under or the scene that pedestrian behavior is unusual intensive the pedestrian, be difficult to pedestrian's agglomerate is effectively cut apart or number is estimated, have a strong impact on the accuracy rate of algorithm.(3) some pedestrian detection, track algorithm have proposed requirement for the resolution of video image, but in real world applications, for the consideration to equipment cost, the resolution of video image is generally on the low side.(4) existing many features passenger flow statistical method, equipment cost is higher, is difficult to apply.
Summary of the invention
The purpose of this invention is to provide a kind of pedestrian's computing method based on many features, solve defective and deficiency that prior art exists.
The invention provides a kind of pedestrian's computing method based on many features, may further comprise the steps:
1.1 above the target area, arrange in order to the infrared transmitter of launching the infrared light spot dot matrix and in order to catch the camera of infrared image;
Be transmitted in hot spot dot matrix in the unmanned described target area 1.2 obtain described infrared transmitter by described camera, set up the background lattice model that comprises described facula position;
Be transmitted in hot spot display in the described target area as the prospect dot matrix 1.3 obtain described infrared transmitter by described camera, facula position in prospect dot matrix and the described background lattice model is compared, the hot spot that the position is changed carries out cluster and extracts with the form of agglomerate, to set up prospect facula deviation model;
Facula deviation local maximum point zone is the head agglomerate as pedestrian's head zone with its cluster in the described prospect facula deviation model 1.4 extract;
1.5 follow the tracks of the motion of described head agglomerate, determine number based on described motion characteristics.
Described step 1.2 specifically may further comprise the steps:
To unmanned scene light spot image, to use self-adaption thresholding method that hot spot is extracted from scene image, and extract light spot profile on this basis, profile center, location obtains first image;
In described first image, obtain hot spot spacing, lattice structure and hot spot neighborhood scope, find out central point;
Central point with described first image is starting point, accurately locatees in the hot spot field, obtains described background lattice model.
Described step 1.3 specifically may further comprise the steps:
Upgrade the facula position of current frame image, compare with background model, extract the zone that skew takes place hot spot, obtain second image;
Use neighbour's interpolation algorithm to fill up facula position in the facula deviation zone to described second image, reduce hot spot for the influence of pedestrian's feature point extraction, obtain the 3rd image;
Hot spot distance and unit distance off-set value according to described the 3rd image are carried out cluster to the prospect facula position, extract prospect with the form of agglomerate.
Further comprising the steps of after described step 1.4 is finished:
False head is rejected: reject the false head zone that comprises shoulder or noise;
Pedestrian's identification: all the other skew hot spots are included into each head zone according to distance and unit distance increasing degree.
Described step 1.5 specifically may further comprise the steps:
Optical flow computation: the LK light stream of calculating sequential frame image;
Two-way coupling: based on the interframe light stream pedestrian's target is carried out two-way coupling, according to one to one, merge, three kinds of schema update pursuit paths of division, and further revise goal recognition result on this basis;
Demographics: based on pursuit path number is added up, use two detection line mark scene areas, carry out demographics when the pedestrian passes through two detection lines successively, the order that the pedestrian passes through two detection lines is considered as pedestrian's direct of travel.
Before carrying out described optical flow computation, also comprise and carry out the step that the image histogram equalization is handled.
Camera in the described step 1.1 is the ccd video camera that installs infrared filter membrane additional.
The present invention is by setting up background lattice model and the prospect of seizure dot matrix, both facula positions are compared and set up prospect facula deviation model, extracting the regional head zone as the pedestrian of facula deviation local maximum point wherein again, is the head agglomerate with its cluster; Last pursuit movement track is determined statistical number of person based on described motion characteristics.The present invention is based on optical flow computation, use the two-way coupling of interframe agglomerate to realize that the individual track of pedestrian upgrades, demographics is precisely reliable.On the other hand, the present invention compares with the method based on the dense degree of depth also based on sparse depth data, has improved the execution speed of algorithm.At last, compare with data acquisition equipment such as TOF camera, it is cheap that the relative depth that the present invention adopts is obtained equipment cost, has bigger advantage at cost.
Description of drawings
Fig. 1 is schematic flow sheet of the present invention;
Fig. 2 is the schematic flow sheet of one embodiment of the invention;
Fig. 3 is the unmanned scene synoptic diagram in the background modeling step;
Fig. 4 is the adaptive threshold synoptic diagram in the background modeling step;
Fig. 5 is that profile centralized positioning and the array structure in the background modeling step obtains synoptic diagram;
Fig. 6 is location, the tabula rasa position synoptic diagram in the background modeling step;
Fig. 7 is the background lattice model synoptic diagram in the background modeling step;
Fig. 8 is the two field picture synoptic diagram in the prospect modeling procedure;
Fig. 9 is the renewal facula position synoptic diagram in the prospect modeling procedure;
Figure 10 is the facula position skew synoptic diagram in the prospect modeling procedure;
Figure 11 is the skew hot spot interpolation synoptic diagram in the prospect modeling procedure;
Figure 12 is the agglomerate cluster synoptic diagram in the prospect modeling procedure;
Figure 13 is the former scene image synoptic diagram in the count tracking step;
Figure 14 is the histogram equalization synoptic diagram in the count tracking step;
Figure 15 is the calculating light stream synoptic diagram in the count tracking step;
Figure 16 is the two-way coupling synoptic diagram of the interframe in the count tracking step;
Figure 17 is the count tracking result schematic diagram in the count tracking step.
Embodiment
Further specify technical scheme of the present invention below in conjunction with drawings and embodiments.
Referring to Fig. 1, the invention provides a kind of pedestrian's computing method based on many features, may further comprise the steps:
1.1 above the target area, arrange in order to the infrared transmitter of launching the infrared light spot dot matrix and in order to catch the camera of infrared image.
As an embodiment, the camera in the described step 1.1 is the ccd video camera that installs infrared filter membrane additional, can catch infrared image.
Be transmitted in hot spot dot matrix in the unmanned described target area 1.2 obtain described infrared transmitter by described camera, set up the background lattice model that comprises described facula position.
As an embodiment, this step 1.2 is specifically further comprising the steps of: to the unmanned scene light spot image among Fig. 3, use self-adaption thresholding method that hot spot is extracted from scene image, and extract light spot profile on this basis, profile center, location obtains first image, referring to Fig. 4.Referring to Fig. 5, in described first image, obtain hot spot spacing, lattice structure and hot spot neighborhood scope, find out central point, be starting point with the central point of described first image, in the hot spot field, accurately locate (Fig. 6), obtain described background dot matrix illustraton of model 7.
Be transmitted in hot spot display in the described target area as the prospect dot matrix 1.3 obtain described infrared transmitter by described camera, facula position in prospect dot matrix and the described background lattice model is compared, the hot spot that the position is changed carries out cluster and extracts with the form of agglomerate, to set up prospect facula deviation model.
As an embodiment, described step 1.3 specifically may further comprise the steps: referring to Fig. 8, Fig. 9, upgrade the facula position of current frame image, compare with background model, extract the zone that skew takes place hot spot, obtain second image (Figure 10); Use neighbour's interpolation algorithm to fill up facula position in the facula deviation zone to described second image, reduce hot spot for the influence of pedestrian's feature point extraction, obtain the 3rd image (Figure 11); Hot spot distance and unit distance off-set value according to described the 3rd image are carried out cluster to the prospect facula position, extract prospect (Figure 12) with the form of agglomerate.
Facula deviation local maximum point zone is the head agglomerate as pedestrian's head zone with its cluster in the described prospect facula deviation model 1.4 extract.
As an embodiment, further comprising the steps of after described step 1.4 is finished: false head is rejected: reject the false head zone that comprises shoulder or noise; Pedestrian's identification: all the other skew hot spots are included into each head zone according to distance and unit distance increasing degree.
1.5 follow the tracks of the motion of described head agglomerate, determine number based on described motion characteristics.
As an embodiment, described step 1.5 specifically may further comprise the steps: referring to Figure 13 and Figure 14, image is carried out the image histogram equalization handle.Because infrared image is compared with normal image, there are shortcomings such as the target-to-background contrast is poor, object edge fuzzy, noise is big mostly, need carry out histogram equalization to infrared image and handle.Referring to Figure 15, optical flow computation: the LK light stream of calculating sequential frame image.Referring to Figure 16, two-way coupling: based on the interframe light stream pedestrian's target is carried out two-way coupling, according to one to one, merge, three kinds of schema update pursuit paths of division, and further revise goal recognition result on this basis.Referring to Figure 17, demographics: based on pursuit path number is added up, use two detection line mark scene areas, carry out demographics when the pedestrian passes through two detection lines successively, the order that the pedestrian passes through two detection lines is considered as pedestrian's direct of travel.
Participate in Fig. 2, Fig. 2 is the schematic flow sheet of an embodiment, at first read frame of video and judge frame headed by it whether, if then carry out background modeling (narrating above the concrete steps), then carry out the tabula rasa position if not and upgrade to carry out prospect modeling (narrating above the concrete steps), extract pedestrian head zone, optical flow computation, two-way coupling, demographics etc. according to top order successively again.The method of the invention mainly is made of foreground detection, target identification and three modules of count tracking, and its treatment scheme as mentioned above.First, the foreground detection module, light spot image based on unmanned scene carries out background modeling, according to the hot spot structure scene is divided into little hot spot neighborhood, in the zonule, obtain the position of hot spot in image, finish the detection of algorithm applicability according to the power that is detected as of facula position in the background modeling again.Then extract the change in location of hot spot in current scene and the background model corresponding region, recover the relative depth information of region-of-interest, the spot area that the polymerization degree of depth changes is determined sport foreground.Second portion, the target identification module, from prospect, seek the head zone of local maximum representative in the relative depth data of region-of-interest, and other region-of-interests in the prospect agglomerate are arrived each head zone according to regular cluster, thereby pedestrian's agglomerate is effectively cut apart.Third part, pedestrian's counting module, the two-way coupling of using pyramid LK optical flow algorithm to carry out between the frame of front and back in conjunction with the infrared video half-tone information realizes that the pedestrian follows the tracks of pedestrian's number statistics of going forward side by side.
Among the present invention, propose a kind of relative depth and obtained model, this model is based on infrared external reflection model and camera model, and the reflection of infrared beam highly is presented as the off-set value of corresponding region hot spot in the image, and the facula deviation measurer of the diverse location in the whole scene has certain comparability.The ccd video camera that the logical filter membrane of band has been installed in use catches infrared laser at the light spot image of specifying the no man's land to form.Use adaptive threshold and edge detection algorithm to obtain facula position, carry out background modeling.Extract facula position in the light spot image of current scene, compare with corresponding facula position in the background model, the region-of-interest that the polymerization position changes constitutes the prospect agglomerate.The off-set value of the same area hot spot is this regional relative depth in prospect dot matrix and the background dot matrix.In the prospect agglomerate, seek the local maximum of relative depth, be designated potential head zone; In potential head zone, reject as shoulder or the false head zone that caused by noise; Other prospect spot area of polymerization constitute the target agglomerate.Use pyramid LK optical flow algorithm to carry out target following.Between the video sequence successive frame, use two-way coupling to finish the renewal of target following track, play certain agglomerate simultaneously and cut apart correcting action.
Those of ordinary skill in the art will be appreciated that, above embodiment only is that the present invention is described, and be not to be used as limitation of the invention, as long as in connotation scope of the present invention, all will drop in claims scope of the present invention variation, the modification of above embodiment.
Claims (7)
1. the pedestrian's computing method based on many features is characterized in that, may further comprise the steps:
1.1 above the target area, arrange in order to the infrared transmitter of launching the infrared light spot dot matrix and in order to catch the camera of infrared image;
Be transmitted in hot spot dot matrix in the unmanned described target area 1.2 obtain described infrared transmitter by described camera, set up the background lattice model that comprises described facula position;
Be transmitted in hot spot display in the described target area as the prospect dot matrix 1.3 obtain described infrared transmitter by described camera, facula position in prospect dot matrix and the described background lattice model is compared, the hot spot that the position is changed carries out cluster and extracts with the form of agglomerate, to set up prospect facula deviation model;
Facula deviation local maximum point zone is the head agglomerate as pedestrian's head zone with its cluster in the described prospect facula deviation model 1.4 extract;
1.5 follow the tracks of the motion of described head agglomerate, determine number based on described motion characteristics.
2. the method for claim 1 is characterized in that, described step 1.2 specifically may further comprise the steps:
To unmanned scene light spot image, to use self-adaption thresholding method that hot spot is extracted from scene image, and extract light spot profile on this basis, profile center, location obtains first image;
In described first image, obtain hot spot spacing, lattice structure and hot spot neighborhood scope, find out central point;
Central point with described first image is starting point, accurately locatees in the hot spot field, obtains described background lattice model.
3. method as claimed in claim 1 or 2 is characterized in that, described step 1.3 specifically may further comprise the steps:
Upgrade the facula position of current frame image, compare with background model, extract the zone that skew takes place hot spot, obtain second image;
Use neighbour's interpolation algorithm to fill up facula position in the facula deviation zone to described second image, reduce hot spot for the influence of pedestrian's feature point extraction, obtain the 3rd image;
Hot spot distance and unit distance off-set value according to described the 3rd image are carried out cluster to the prospect facula position, extract prospect with the form of agglomerate.
4. method as claimed in claim 3 is characterized in that, and is further comprising the steps of after described step 1.4 is finished:
False head is rejected: reject the false head zone that comprises shoulder or noise;
Pedestrian's identification: all the other skew hot spots are included into each head zone according to distance and unit distance increasing degree.
5. method as claimed in claim 4 is characterized in that, described step 1.5 specifically may further comprise the steps:
Optical flow computation: the LK light stream of calculating sequential frame image;
Two-way coupling: based on the interframe light stream pedestrian's target is carried out two-way coupling, according to one to one, merge, three kinds of schema update pursuit paths of division, and further revise goal recognition result on this basis;
Demographics: based on pursuit path number is added up, use two detection line mark scene areas, carry out demographics when the pedestrian passes through two detection lines successively, the order that the pedestrian passes through two detection lines is considered as pedestrian's direct of travel.
6. method as claimed in claim 5 is characterized in that, also comprises carrying out the step that the image histogram equalization is handled before carrying out described optical flow computation.
7. method as claimed in claim 6 is characterized in that, the camera in the described step 1.1 is the ccd video camera that installs infrared filter membrane additional.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310211139.6A CN103279791B (en) | 2013-05-30 | 2013-05-30 | Based on pedestrian's computing method of multiple features |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310211139.6A CN103279791B (en) | 2013-05-30 | 2013-05-30 | Based on pedestrian's computing method of multiple features |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103279791A true CN103279791A (en) | 2013-09-04 |
| CN103279791B CN103279791B (en) | 2016-03-30 |
Family
ID=49062303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310211139.6A Active CN103279791B (en) | 2013-05-30 | 2013-05-30 | Based on pedestrian's computing method of multiple features |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103279791B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104394378A (en) * | 2014-12-05 | 2015-03-04 | 成都美联微智科技有限公司 | Novel monitoring device for tourism area passenger flow analysis |
| CN104573613A (en) * | 2013-10-16 | 2015-04-29 | 深圳市捷顺科技实业股份有限公司 | Video security anti-smashing method and device based on blob tracking |
| CN104908048A (en) * | 2014-03-14 | 2015-09-16 | 本田技研工业株式会社 | Motion estimation device, robot, and motion estimation method |
| CN105446351A (en) * | 2015-11-16 | 2016-03-30 | 杭州码全信息科技有限公司 | Robotic airship system capable of locking target area for observation based on autonomous navigation |
| CN106446788A (en) * | 2016-08-31 | 2017-02-22 | 山东恒宇电子有限公司 | Method for passenger flow statistic by means of high-dynamic range image based on optic nerve mechanism |
| CN106600777A (en) * | 2016-12-09 | 2017-04-26 | 济南赛英立德电子科技有限公司 | Infrared array number-of-personnel sensor-based counting method and apparatus |
| CN106846297A (en) * | 2016-12-21 | 2017-06-13 | 深圳市镭神智能***有限公司 | Pedestrian's flow quantity detecting system and method based on laser radar |
| CN107221175A (en) * | 2017-05-31 | 2017-09-29 | 深圳市鸿逸达科技有限公司 | A kind of pedestrian is intended to detection method and system |
| CN107408205A (en) * | 2015-03-11 | 2017-11-28 | 微软技术许可有限责任公司 | Foreground and background is distinguished with infrared imaging |
| CN109076190A (en) * | 2016-04-22 | 2018-12-21 | 株式会社S1 | Detect the device and method of abnormal conditions |
| CN109359552A (en) * | 2018-09-21 | 2019-02-19 | 中山大学 | One kind is efficiently across camera shooting head's bi-directional tracking method |
| CN109447016A (en) * | 2018-11-07 | 2019-03-08 | 浩云科技股份有限公司 | A kind of demographic method and system between adding paper money based on structure light |
| CN111291736A (en) * | 2020-05-07 | 2020-06-16 | 南京景三医疗科技有限公司 | Image correction method and device and medical equipment |
| CN112232210A (en) * | 2020-10-16 | 2021-01-15 | 京东方科技集团股份有限公司 | Personnel flow analysis method and system, electronic device and readable storage medium |
| CN114241015A (en) * | 2022-02-28 | 2022-03-25 | 长沙莫之比智能科技有限公司 | Method and device for counting number of people going in and out based on millimeter wave radar |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101826228A (en) * | 2010-05-14 | 2010-09-08 | 上海理工大学 | Detection method of bus passenger moving objects based on background estimation |
| CN102592335A (en) * | 2010-12-14 | 2012-07-18 | 施乐公司 | Determining a total number of people in an ir image obtained via an ir imaging system |
| KR20130021596A (en) * | 2011-08-23 | 2013-03-06 | 리누딕스 주식회사 | Apparatus for sensing human body with improved accuracy |
-
2013
- 2013-05-30 CN CN201310211139.6A patent/CN103279791B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101826228A (en) * | 2010-05-14 | 2010-09-08 | 上海理工大学 | Detection method of bus passenger moving objects based on background estimation |
| CN102592335A (en) * | 2010-12-14 | 2012-07-18 | 施乐公司 | Determining a total number of people in an ir image obtained via an ir imaging system |
| KR20130021596A (en) * | 2011-08-23 | 2013-03-06 | 리누딕스 주식회사 | Apparatus for sensing human body with improved accuracy |
Non-Patent Citations (2)
| Title |
|---|
| GWANG-GOOK LEE ETAL.: "Pedestrian Counting using an IR Line Laser", 《INTERNATIONAL CONFERENCE ON CONVERGENCE AND HYBRID INFORMATION TECHNOLOGY 2008》 * |
| 林宇: "基于红外激光线检测的乘客计数方法", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573613A (en) * | 2013-10-16 | 2015-04-29 | 深圳市捷顺科技实业股份有限公司 | Video security anti-smashing method and device based on blob tracking |
| CN104573613B (en) * | 2013-10-16 | 2018-05-01 | 深圳市捷顺科技实业股份有限公司 | A kind of Video security based on mass tracking is prevented pounding method and device |
| CN104908048A (en) * | 2014-03-14 | 2015-09-16 | 本田技研工业株式会社 | Motion estimation device, robot, and motion estimation method |
| US9684824B2 (en) | 2014-03-14 | 2017-06-20 | Honda Motor Co., Ltd | Motion estimation device, robot, and method to estimate reaction of a person |
| CN104394378A (en) * | 2014-12-05 | 2015-03-04 | 成都美联微智科技有限公司 | Novel monitoring device for tourism area passenger flow analysis |
| CN107408205A (en) * | 2015-03-11 | 2017-11-28 | 微软技术许可有限责任公司 | Foreground and background is distinguished with infrared imaging |
| CN105446351A (en) * | 2015-11-16 | 2016-03-30 | 杭州码全信息科技有限公司 | Robotic airship system capable of locking target area for observation based on autonomous navigation |
| CN109076190A (en) * | 2016-04-22 | 2018-12-21 | 株式会社S1 | Detect the device and method of abnormal conditions |
| CN109076190B (en) * | 2016-04-22 | 2021-03-02 | 株式会社S1 | Apparatus and method for detecting abnormal condition |
| US10853664B2 (en) | 2016-04-22 | 2020-12-01 | S-1 Corporation | Device and method for detecting abnormal situation |
| CN106446788A (en) * | 2016-08-31 | 2017-02-22 | 山东恒宇电子有限公司 | Method for passenger flow statistic by means of high-dynamic range image based on optic nerve mechanism |
| CN106600777A (en) * | 2016-12-09 | 2017-04-26 | 济南赛英立德电子科技有限公司 | Infrared array number-of-personnel sensor-based counting method and apparatus |
| CN106600777B (en) * | 2016-12-09 | 2019-02-22 | 济南赛英立德电子信息股份有限公司 | Method of counting and device based on infrared array number sensor |
| CN106846297A (en) * | 2016-12-21 | 2017-06-13 | 深圳市镭神智能***有限公司 | Pedestrian's flow quantity detecting system and method based on laser radar |
| CN107221175A (en) * | 2017-05-31 | 2017-09-29 | 深圳市鸿逸达科技有限公司 | A kind of pedestrian is intended to detection method and system |
| CN109359552A (en) * | 2018-09-21 | 2019-02-19 | 中山大学 | One kind is efficiently across camera shooting head's bi-directional tracking method |
| CN109359552B (en) * | 2018-09-21 | 2020-11-13 | 中山大学 | Efficient cross-camera bidirectional pedestrian tracking method |
| CN109447016A (en) * | 2018-11-07 | 2019-03-08 | 浩云科技股份有限公司 | A kind of demographic method and system between adding paper money based on structure light |
| CN111291736A (en) * | 2020-05-07 | 2020-06-16 | 南京景三医疗科技有限公司 | Image correction method and device and medical equipment |
| CN112232210A (en) * | 2020-10-16 | 2021-01-15 | 京东方科技集团股份有限公司 | Personnel flow analysis method and system, electronic device and readable storage medium |
| CN114241015A (en) * | 2022-02-28 | 2022-03-25 | 长沙莫之比智能科技有限公司 | Method and device for counting number of people going in and out based on millimeter wave radar |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103279791B (en) | 2016-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103279791B (en) | Based on pedestrian's computing method of multiple features | |
| CN106650620B (en) | A kind of target person identification method for tracing using unmanned plane monitoring | |
| CN107253485B (en) | Foreign matter invades detection method and foreign matter invades detection device | |
| CN101950426B (en) | Vehicle relay tracking method in multi-camera scene | |
| CN103473554B (en) | Artificial abortion's statistical system and method | |
| JP6448223B2 (en) | Image recognition system, image recognition apparatus, image recognition method, and computer program | |
| CN106203513B (en) | A kind of statistical method based on pedestrian's head and shoulder multi-target detection and tracking | |
| CN104978567B (en) | Vehicle checking method based on scene classification | |
| WO2018119668A1 (en) | Method and system for recognizing head of pedestrian | |
| CN106127137A (en) | A kind of target detection recognizer based on 3D trajectory analysis | |
| CN103325115B (en) | A kind of method of monitoring people counting based on overhead camera head | |
| CN108986158A (en) | A kind of across the scene method for tracing identified again based on target and device and Computer Vision Platform | |
| CN102521565A (en) | Garment identification method and system for low-resolution video | |
| CN109344690B (en) | People counting method based on depth camera | |
| CN108776974B (en) | A kind of real-time modeling method method suitable for public transport scene | |
| CN104517095B (en) | A kind of number of people dividing method based on depth image | |
| CN102243765A (en) | Multi-camera-based multi-objective positioning tracking method and system | |
| CN103778645B (en) | Circular target real-time tracking method based on images | |
| CN102043953A (en) | Real-time-robust pedestrian detection method aiming at specific scene | |
| CN104318263A (en) | Real-time high-precision people stream counting method | |
| WO2018076392A1 (en) | Pedestrian statistical method and apparatus based on recognition of parietal region of human body | |
| CN103049758A (en) | Method for realizing remote authentication by fusing gait flow images (GFI) and head and shoulder procrustes mean shapes (HS-PMS) | |
| CN105160297A (en) | Masked man event automatic detection method based on skin color characteristics | |
| CN105160649A (en) | Multi-target tracking method and system based on kernel function unsupervised clustering | |
| CN103593679A (en) | Visual human-hand tracking method based on online machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C53 | Correction of patent of invention or patent application | ||
| CB02 | Change of applicant information |
Address after: 201505 Jinshan District, Shanghai, Lin Ting Town, Temple Ping Road, room 16, No. 599 Applicant after: SHANGHAI WINNER INFORMATION TECHNOLOGY CO., INC. Address before: 201505 Jinshan District, Shanghai, Lin Ting Town, Temple Ping Road, room 16, No. 599 Applicant before: Huina Network Information Science & Technology Co., Ltd. |
|
| COR | Change of bibliographic data |
Free format text: CORRECT: APPLICANT; FROM: HUINA NETWORK INFORMATION SCIENCE + TECHNOLOGY CO., LTD. TO: SHANGHAI WINNER INFORMATION TECHNOLOGY CO., INC. |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 201505 Shanghai city Jinshan District ting Lin Zhen Ting Feng Road No. 333, room 216 Patentee after: Winner Technology Co.,Ltd. Address before: 201505, room 16, No. 599, Ping Nan Road, ting Lin Town, Shanghai, Jinshan District Patentee before: SHANGHAI WINNER INFORMATION TECHNOLOGY Co.,Inc. |
|
| CP02 | Change in the address of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: 201203 No. 6, Lane 55, Chuanhe Road, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Patentee after: Winner Technology Co.,Ltd. Address before: 201505 Room 216, 333 Tingfeng Highway, Tinglin Town, Jinshan District, Shanghai Patentee before: Winner Technology Co.,Ltd. |