CN109063578B - Crowd gathering detection method - Google Patents
Crowd gathering detection method Download PDFInfo
- Publication number
- CN109063578B CN109063578B CN201810734133.XA CN201810734133A CN109063578B CN 109063578 B CN109063578 B CN 109063578B CN 201810734133 A CN201810734133 A CN 201810734133A CN 109063578 B CN109063578 B CN 109063578B
- Authority
- CN
- China
- Prior art keywords
- crowd
- target
- area
- situation
- direction points
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
- G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
Abstract
The invention relates to the field of intelligent video monitoring, in particular to a crowd gathering detection method. Firstly, carrying out multi-target intelligent identification and tracking to obtain a target, a target coordinate, a target moving speed and a target moving direction in a video image; secondly, calculating according to the target and the target coordinate to obtain the local density of the target; automatically constructing a crowd concentration circle according to the target local density and a preset crowd quantity threshold value in the crowd concentration circle; and finally, judging the situation of the crowd enclosure according to the crowd enclosure and the moving direction of the target.
Description
Technical Field
The invention relates to the field of intelligent video monitoring, in particular to a crowd gathering detection method.
Background
With the continuous advance of modernization and urbanization, the population of cities and towns is increasing, and a large number of people often appear in business centers, public transportation, entertainment places, sports places, government doorways, enterprise doorways and the like of cities and towns, so that the frequency of accidents or disasters caused by the frequent occurrence of the cities and the towns is also increasing.
In the face of the situation, how to insist on people-oriented behavior, how to early warn crowd gathering phenomenon through video, and how to effectively reduce accidents or disasters are a great research subject. In order to early warn crowd gathering events, an intelligent video monitoring technology is introduced. However, the traditional video monitoring adopts a large monitoring screen in a large scale, which not only consumes manpower and material resources, but also is easy to miss the occurring colony aggregation events. Meanwhile, after people stare at the monitoring large screen for a long time, visual fatigue is easily caused, and the attention is reduced. With the rapid development and application of high-performance computing and artificial intelligence technologies, people can be quickly warned of crowd gathering events by replacing human eyes with intelligent video analysis technologies.
At present, the traditional crowd detection algorithm based on crowd trajectory, speed, texture information or pixel statistics mainly has the following defects: (1) without setting a fixed area, two or more crowd gathers cannot be detected simultaneously. (2) The situation of the moving direction of the crowd tuck cannot be accurately judged.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a crowd gathering detection method, which obtains crowd gathering circles and judges crowd gathering situations through a multi-target intelligent identification and tracking technology and an advanced clustering algorithm.
In order to achieve the above object, the present invention provides a method for detecting crowd accumulation, comprising the following steps:
a method of crowd detection, the method comprising the steps of:
s1: acquiring a monitoring video stream of a monitored place, and taking each continuous frame image in the video stream as an analysis processing image;
s2: the image is processed by advanced machine learning algorithm, the target in the image is identified and the target coordinate is calculated,for a corresponding target coordinate point data set, where N is the target number;
S3: according to the data set D obtained in step S2, the distance D between each two target points in D is first calculatedij=dist(Pi,Pj) Dist denotes a coordinate point PiAnd PjA certain distance therebetween; then calculating the density of each target coordinate pointWherein the functiondcAnd k mR, wherein R is the number of pixels occupied by a single target, k belongs to (0, 1), and m is a set crowd number threshold in a crowd clustering circle.
S4: according to the target and the target coordinates obtained in the step S2, performing multi-target tracking on each frame of image by a tracking algorithm, calculating to obtain the moving speed of each target and calculating the moving direction;
s5: density ρ obtained in step S3iFirst, calculateφiFor describing the minimum distance from point i to other higher density points, those having a relatively large density piAnd a large phiiIs considered to be the center of the crowd;
then calculating the distance from the target point to the central pointFinally obtaining the crowd concentrationIf it is notA crowd-sourcing circle is determined;
s6: setting the rated direction threshold value of crowd gathering as | HTObtaining crowd concentration according to each target moving direction obtained in step S4 and crowd concentration circle obtained in step S5Loop direction Hi。
Further, calculate HiThe angle value θ and divides the space into the following four regions:
if 0 is more than or equal to theta and less than 90 DEG, then HiThe direction points to the I area;
if the angle is more than or equal to 90 and less than 180 degrees, then HiThe direction points to the area II;
if the angle is more than or equal to 180 degrees and is less than 270 degrees, H isiThe direction points to the III area;
if theta is more than or equal to 270 degrees and less than 360 degrees, H isiThe direction points to the IV area.
According to direction HiAnd densityThe situation of the crowd gathering circle is judged, and the situation is divided into the following eleven situations:
if it is usedBecome smaller, | HiIf | < epsilon, the situation of the crowd gathering circle is dissipation;
if it is notIncreasing the size, | Hi | < epsilon, and the situation of crowd gathering circles is cohesion;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the I area, and the situation of the crowd tucking is that the crowd tucking moves to the I area at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the area II, and the situation of the crowd tucking is that the crowd tucking moves to the area II at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the III area, and the situation of the crowd colony moves to the III area at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to an IV area, and the situation of the crowd tuck is that the crowd tuck moves to the IV area at a low speed;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area I, and the situation of the crowd tucking is that the crowd tucking moves to the area I quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area II, and the situation of the crowd colony is that the crowd colony moves to the area II quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area III, and the situation of the crowd colony is that the crowd colony moves to the area III quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the IV area, and the situation of the crowd ring is that the crowd ring moves to the IV area rapidly.
Further, where ε is a minimum value.
Further, in step S2: and identifying the target in the image and calculating to obtain a target coordinate, wherein the target is a person.
Further, in step S4: the tracking algorithm is one of Kalman Filter or KCF tracking algorithm.
The invention provides a crowd gathering detection method, which comprises the steps of firstly carrying out multi-target intelligent identification and tracking to obtain a target (person), a target coordinate, a target moving speed and a target moving direction in a video image; secondly, calculating according to the target and the target coordinates to obtain the local density of the target; automatically constructing a crowd concentration circle according to the target local density and a preset crowd quantity threshold value in the crowd concentration circle; and finally, judging the situation of the crowd enclosure according to the crowd enclosure and the moving direction of the target.
The invention has the following beneficial effects:
1. a fixed area does not need to be preset to detect crowd gathering;
2. the method is not limited to special scenes, can be used in the daytime, can be used at night, can be used in commercial centers, can be used in public transportation and the like;
3. the location of one or more people groups may be automatically detected;
4. and automatically judging the moving direction situation of the colony of each people by combining the center of each people and the moving direction of the target.
Drawings
FIG. 1 is a flow chart of a method for crowd detection according to the present invention;
FIG. 2 is a schematic flow chart of crowd-sourcing and location calculation;
fig. 3 is a flowchart illustrating the determination of the situation of the crowd zone according to an embodiment.
Detailed Description
The invention is explained in further detail below with reference to the drawings in which:
fig. 1 is a flow chart of a crowd gathering detection method provided by the present invention. As shown in fig. 1, a crowd gathering detection method according to an embodiment of the present invention includes the following steps:
s101: acquiring a detection video stream:
and acquiring a monitoring video stream of the monitored place, and taking each continuous frame image in the video stream as an analysis processing image.
S102: identifying the target and calculating the target coordinates:
processing the images by advanced machine learning algorithm according to the continuous images acquired in step S101, recognizing the target (person) in the images and calculating the coordinates of the target,is a corresponding target coordinate point data set, where N is the target number.
S103: calculating a target local density according to the target and the target coordinates:
according to the data set D obtained in the step S102, firstly, the distance D between every two target points in the D is calculatedij=dist(Pi,Pj) Dist denotes a coordinate point PiAnd PjA certain distance therebetween; then calculating the density of each target coordinate pointWherein the functiondcAnd k mR, wherein R is the number of pixels occupied by a single target, k belongs to (0, 1), and m is a set crowd number threshold in a crowd clustering circle.
S104: target tracking and target moving direction obtaining:
according to the target and target coordinates obtained in step S102, performing multi-target tracking on each frame image by a tracking algorithm (such as Kalman Filter, KCF and other tracking algorithms) and calculating to obtain the moving speed v of each targetiThen calculate the ith target direction hi=Pk(x,y)-Pi(x, y) wherein Pi(x, y) is the ith target coordinate, Pk(X)=Pi(X)+αvi,Pk(y)=Pi(y)+avi。
S105: automatically constructing a crowd-together circle:
next, the target local density obtained in step S103 is analyzed and calculated to automatically construct a crowd coil. Fig. 2 is a flow diagram of crowd tucking and position calculation. As shown in fig. 2, the crowd-sourcing and location calculation employed in the present invention includes the following steps:
s201: calculating a metric value phii:
Density ρ of target coordinate pointiCalculatingφiFor describing the minimum distance from point i to other higher density points, those having a relatively large density piAnd a large phiiIs considered to be the center of the crowd.
S202: calculating the distance from the target point to the central point:
according to the measured value phiiCalculating the distance from each target point to the central point of each crowd concentration
S203: obtaining the clustering circle and position of each person:
according toObtaining the concentration of the crowdIf it is usedThen a people cluster is determined and so on all people cluster and cluster center positions are obtained.
S106: judging the situation of the crowd accumulation ring according to the crowd accumulation ring and the target moving direction:
next, the situation of the crowd at the position of each crowd at the target moving direction obtained in step S104 and the situation of the crowd at the position of each crowd at step S105 are determined. Fig. 3 is a flowchart illustrating the determination of the situation of the crowd zone according to an embodiment. As shown in fig. 3, the method for determining the crowd accumulation situation in the present invention comprises the following steps:
s301: obtaining the direction H of the crowd zonei:
Setting the rated direction threshold value of crowd gathering as | HTL, obtaining the direction H of the crowd 'S gather area according to the moving speed and direction of each target obtained in step S4 and the crowd' S gather area obtained in step S105i,HiFor each target direction hiThe resulting vectors are added.
S302: the space is divided into four regions:
calculate HiThe angle value θ and divides the space into the following four regions:
if 0 is more than or equal to theta is less than 90 DEG, then HiThe direction points to the I area;
if the angle is more than or equal to 90 and less than 180 degrees, then HiThe direction points to the area II;
if the angle is more than or equal to 180 degrees and is less than 270 degrees, H isiThe direction points to the III area;
if theta is more than or equal to 270 degrees and less than 360 degrees, H isiThe direction points to the IV area.
S303: judging the situation of the crowd gathering circle:
according to direction HiAnd densityThe situation of the crowd gathering circle is judged, and the situation is divided into the following eleven situations:
if it is notBecome smaller, | HiIf | < epsilon, the situation of the crowd gathering circle is dissipation;
if it is notBecome large, | HiIf | < epsilon, the situation of the crowd gathering circles is cohesion;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the I area, and the situation of the crowd tucking is that the crowd tucking moves to the I area at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the area II, and the situation of the crowd tucking is that the crowd tucking moves to the area II at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the area III, and the situation of the crowd gathering circle is that the crowd gathers people to move to the area III at a low speed;
if it is notInvariably,. epsilon. < | HiThe I < | HT | points to the IV area in the Hi direction, and the situation of crowd tucking is that people move to the IV area at a low speed;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area I, and the situation of the crowd tucking is that the crowd tucking moves to the area I quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area II, and the situation of the crowd colony is that the crowd colony moves to the area II quickly;
if it is usedInvariable, | HT|<|Hi|,HiThe direction points to the area III, and the situation of the crowd colony is that the crowd colony moves to the area III quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the IV area, and the situation of the crowd ring is that the crowd ring moves to the IV area rapidly.
Where ε is a minimum value.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (4)
1. A method of crowd detection, the method comprising the steps of:
s1: acquiring a monitoring video stream of a monitored place, and taking each continuous frame image in the video stream as an analysis processing image;
s2: the image is processed by advanced machine learning algorithm, the target in the image is identified and the target coordinate is calculated,corresponding target coordinate point data sets are obtained, wherein N is the target number;
s3: according to the data set D obtained in step S2, the distance D between each two target points in D is first calculatedij=dist(Pi,Pj) Dist denotes a coordinate point PiAnd PjA certain distance therebetween; then calculating the density of each target coordinate pointWherein the functiondcThe number of pixels occupied by a single target is determined as k mR, k belongs to (0, 1), and m is a set crowd number threshold in a crowd clustering circle;
s4: according to the target and the target coordinates obtained in the step S2, performing multi-target tracking on each frame of image by a tracking algorithm, calculating to obtain the moving speed of each target and calculating the moving direction;
s5: density ρ obtained in step S3iFirst, calculateφiFor describing the smaller distance between point i to other higher density points, the center of the crowd concentration is given by ρiValue sum phiiValue determination, where the center of population p is aggregatediLarge value of phiiThe value is also large;
then calculating the distance from the target point to the central pointFinally obtaining the crowd concentrationIf it is notA crowd-sourcing circle is determined; according to the target density value obtained in the step S5, the rho value in the crowd gathering circle is obtainediThe highest value is defined as the center density of the crowd zone
S6: setting the rated direction threshold value of crowd gathering as | HTVector addition is carried out according to the moving speed and the moving direction of each target obtained in the step S4, and the crowd gathering circle direction H is obtained by combining the crowd gathering circle obtained in the step S5i。
2. The method of crowd gathering detection as recited in claim 1, further comprising:
calculating HiThe angle value θ and divides the space into the following four regions:
if theta is more than or equal to 0 degrees and less than 90 degrees, then HiThe direction points to the I area;
if theta is more than or equal to 90 degrees and less than 180 degrees, then HiThe direction points to the area II;
if theta is more than or equal to 180 degrees and less than 270 degrees, then HiThe direction points to the III area;
if theta is more than or equal to 270 degrees and less than 360 degrees, then HiThe direction points to the IV area;
according to direction HiAnd densityThe situation of the crowd gathering circle is judged, and the situation is divided into the following eleven situations:
if it is notBecome smaller, | HiIf | < epsilon, the situation of the crowd gathering circle is dissipation;
if it is notBecome large, | HiIf | < epsilon, the situation of the crowd gathering circle is cohesion;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the I area, and the situation of the crowd tucking is that the crowd tucking moves to the I area at a low speed;
if it is usedInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the area II, and the situation of the crowd tucking is that the crowd tucking moves to the area II at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to the III area, and the situation of the crowd colony moves to the III area at a low speed;
if it is notInvariably,. epsilon. < | Hi|<|HT|,HiThe direction points to an IV area, and the situation of the crowd tuck is that the crowd tuck moves to the IV area at a low speed;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area I, and the situation of the crowd tucking is that the crowd tucking moves to the area I quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area II, and the situation of the crowd colony is that the crowd colony moves to the area II quickly;
if it is notInvariable, | HT|<|Hi|,HiThe direction points to the area III, and the situation of the crowd colony is that the crowd colony moves to the area III quickly;
3. The method of crowd gathering detection as recited in claim 1, further comprising: in step S2: and identifying the target in the image and calculating to obtain a target coordinate, wherein the target is a person.
4. The method of crowd gathering detection as recited in claim 1, further comprising: in step S4: the tracking algorithm is one of Kalman Filter or KCF tracking algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810734133.XA CN109063578B (en) | 2018-07-06 | 2018-07-06 | Crowd gathering detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810734133.XA CN109063578B (en) | 2018-07-06 | 2018-07-06 | Crowd gathering detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109063578A CN109063578A (en) | 2018-12-21 |
CN109063578B true CN109063578B (en) | 2022-05-06 |
Family
ID=64818806
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810734133.XA Active CN109063578B (en) | 2018-07-06 | 2018-07-06 | Crowd gathering detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109063578B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111680830B (en) * | 2020-05-25 | 2024-01-26 | 广州衡昊数据科技有限公司 | Epidemic situation prevention method and device based on aggregation risk early warning |
CN112287890B (en) * | 2020-11-23 | 2023-08-04 | 杭州海康威视数字技术股份有限公司 | Personnel aggregation detection method, readable storage medium, and electronic device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101325690A (en) * | 2007-06-12 | 2008-12-17 | 上海正电科技发展有限公司 | Method and system for detecting human flow analysis and crowd accumulation process of monitoring video flow |
CN101795395A (en) * | 2009-02-04 | 2010-08-04 | 深圳市先进智能技术研究所 | System and method for monitoring crowd situation |
CN103413321A (en) * | 2013-07-16 | 2013-11-27 | 南京师范大学 | Crowd behavior model analysis and abnormal behavior detection method under geographical environment |
CN104202719A (en) * | 2014-08-14 | 2014-12-10 | 长安通信科技有限责任公司 | People number testing and crowd situation monitoring method and system based on position credibility |
CN104680140A (en) * | 2015-02-13 | 2015-06-03 | 电子科技大学 | Image-based crowd massing state detection method |
CN104933412A (en) * | 2015-06-16 | 2015-09-23 | 电子科技大学 | Abnormal state detection method of medium and high density crowd |
CN107330372A (en) * | 2017-06-05 | 2017-11-07 | 四川大学 | A kind of crowd density based on video and the analysis method of unusual checking system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8195598B2 (en) * | 2007-11-16 | 2012-06-05 | Agilence, Inc. | Method of and system for hierarchical human/crowd behavior detection |
-
2018
- 2018-07-06 CN CN201810734133.XA patent/CN109063578B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101325690A (en) * | 2007-06-12 | 2008-12-17 | 上海正电科技发展有限公司 | Method and system for detecting human flow analysis and crowd accumulation process of monitoring video flow |
CN101795395A (en) * | 2009-02-04 | 2010-08-04 | 深圳市先进智能技术研究所 | System and method for monitoring crowd situation |
CN103413321A (en) * | 2013-07-16 | 2013-11-27 | 南京师范大学 | Crowd behavior model analysis and abnormal behavior detection method under geographical environment |
CN104202719A (en) * | 2014-08-14 | 2014-12-10 | 长安通信科技有限责任公司 | People number testing and crowd situation monitoring method and system based on position credibility |
CN104680140A (en) * | 2015-02-13 | 2015-06-03 | 电子科技大学 | Image-based crowd massing state detection method |
CN104933412A (en) * | 2015-06-16 | 2015-09-23 | 电子科技大学 | Abnormal state detection method of medium and high density crowd |
CN107330372A (en) * | 2017-06-05 | 2017-11-07 | 四川大学 | A kind of crowd density based on video and the analysis method of unusual checking system |
Non-Patent Citations (5)
Title |
---|
Clustering by fast search and find of density peaks;Alex Rodriguez 等;《SCIENCE》;20140627;第344卷(第6191期);第1492-1496页 * |
Coherent Filtering:Detecting Coherent Motions from Crowd Clutters;Bolei Zhou 等;《ECCV 2012》;20121231;第857-871页 * |
Fast training algorithm for feedforward neural networks: application to crowd estimation at underground stations;T.W.S. Chow 等;《Artificial Intelligence in Engineering》;19991231;第13卷(第3期);第301-307页 * |
基于行人占据区域的泰森多边形人员密度统计方法;吴春林 等;《火灾科学》;20180630;第27卷(第2期);第114-123页 * |
大规模人群聚集风险控制态势模型研究;谢智昊等;《中国公共安全·学术版》;20171231(第3期);第29-35页 * |
Also Published As
Publication number | Publication date |
---|---|
CN109063578A (en) | 2018-12-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10121079B2 (en) | Video tracking systems and methods employing cognitive vision | |
Chen et al. | A hierarchical model incorporating segmented regions and pixel descriptors for video background subtraction | |
Denman et al. | Improved simultaneous computation of motion detection and optical flow for object tracking | |
CN110688987A (en) | Pedestrian position detection and tracking method and system | |
CN111260693B (en) | High-altitude parabolic detection method | |
JP2019505866A (en) | Passerby head identification method and system | |
CN107657244B (en) | Human body falling behavior detection system based on multiple cameras and detection method thereof | |
CN108596028B (en) | Abnormal behavior detection algorithm based on video recording | |
JP2013196682A (en) | Group-of-people detection method and group-of-people detector | |
JP6616521B2 (en) | Image processing device | |
CN104954747B (en) | Video monitoring method and device | |
CN109145696B (en) | Old people falling detection method and system based on deep learning | |
CN108875531B (en) | Face detection method, device and system and computer storage medium | |
CN109063578B (en) | Crowd gathering detection method | |
JP2020149642A (en) | Object tracking device and object tracking method | |
Xu et al. | Dynamic obstacle detection based on panoramic vision in the moving state of agricultural machineries | |
CN111797803A (en) | Road guardrail abnormity detection method based on artificial intelligence and image processing | |
CN105469054B (en) | The model building method of normal behaviour and the detection method of abnormal behaviour | |
Xia et al. | Vision-based traffic accident detection using matrix approximation | |
CN106845378B (en) | A method of human body target is identified from opposite image | |
WO2023273010A1 (en) | High-rise littering detection method, apparatus, and device, and computer storage medium | |
CN111627049A (en) | High-altitude parabola determination method and device, storage medium and processor | |
Ashraf et al. | HVD-net: a hybrid vehicle detection network for vision-based vehicle tracking and speed estimation | |
CN110084146B (en) | Pedestrian detection method and device based on shielding perception self-supervision learning | |
WO2020137193A1 (en) | Human detection device and human detection method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |