CN110166744A - A kind of monitoring method violating the regulations of setting up a stall based on video geography fence - Google Patents

A kind of monitoring method violating the regulations of setting up a stall based on video geography fence Download PDF

Info

Publication number
CN110166744A
CN110166744A CN201910350804.7A CN201910350804A CN110166744A CN 110166744 A CN110166744 A CN 110166744A CN 201910350804 A CN201910350804 A CN 201910350804A CN 110166744 A CN110166744 A CN 110166744A
Authority
CN
China
Prior art keywords
video
geography fence
monitoring
regulations
fence
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910350804.7A
Other languages
Chinese (zh)
Inventor
刘学军
倪锡春
王美珍
虞湘
吴春平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Deqing County Geographic Information Center
Nanjing Normal University
Original Assignee
Deqing County Geographic Information Center
Nanjing Normal University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Deqing County Geographic Information Center, Nanjing Normal University filed Critical Deqing County Geographic Information Center
Priority to CN201910350804.7A priority Critical patent/CN110166744A/en
Publication of CN110166744A publication Critical patent/CN110166744A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/50Context or environment of the image
    • G06V20/52Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00Television systems
    • H04N7/18Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/241Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The monitoring method violating the regulations of setting up a stall based on video geography fence that the invention discloses a kind of, comprising the following steps: S1: draw the ruler map of video monitoring regional;S2: establishing video image space and geographical space mapping relations, obtains corresponding homography matrix;S3: geography fence region is established, extracts the fundamental of fence, the fundamental of the fence includes place, time, target and rule;S4: the inverse matrix based on the homography matrix that the fence band of position obtained in S3 and S2 obtain establishes the two-way linkage of sdi video and geography fence;S5: using deep learning method, manages facility's objectives to the flowing in video and identifies;S6: according to the positional relationship between point and polygon, judge that the positional relationship of facility's objectives and geography fence is managed in flowing, determine whether there is violating the regulations set up a stall.

Description

A kind of monitoring method violating the regulations of setting up a stall based on video geography fence
Technical field
The present invention relates to a kind of, and the detection method violating the regulations of setting up a stall based on video geography fence is specifically by video The technical methods such as reason fence, deep learning images steganalysis are identified, positioned and are judged to violating the regulations set up a stall.The present invention relates to City management monitoring technology field.
Background technique
Since 21 century, China is in the social transformation stage sharply, along with the rapid development of current social economy, city The quickening of city's process, city size constantly expand, and urban population is also being continuously increased, citizen's behavioral activity increasingly complex and Diversification.The uncontrollable factor of social human behavior, which increases, brings huge pressure to community service and monitoring control.In public affairs Place altogether, steal plunder, it is violated cross, lane change violating the regulations, occupy phenomena such as public domain in violation of rules and regulations and occur again and again so that security protection, friendship The service in the fields such as logical, city management, supervision early warning become current Chinese society and develop Important Problems urgently to be resolved (the design and realization of urban grid management and service system of Li Deren, Peng Mingjun, the Shao Zhenfeng based on spatial database [J] Wuhan University Journal: information science version, 2006,31 (6): 471-475.).Traditional city management mode under this background Being pushed further into for new period urbanization is not adapted to.The innovation of modern science and technology and information technology develops, and is also city The innovation of city's management mode provides more references and technical support.(Chen Ping Digital Urban Management Model analyses the north [J] Capital college journal: philosophy and the social sciences version, 2006, Vol.43 (1): 142-148;Wang R,Li S.Research about Current Condition and Development of Digital Urban Management in Chongqing [C].International Conference on Control.IEEE,2011.)。
Traditional departmentalization Urban management mode is due to information asymmetry, shortage supervision check-and-balance system, it is difficult to pendulum violating the regulations Booth phenomenon is managed, and main cause has following: 1) such stand mobility is big, without fixed place;2) such to set up a stall The personnel of setting up an office are mostly the self-employed personnel of long campaigns;3) city is only limitted to for this kind of source of deploying to ensure effective monitoring and control of illegal activities set up a stall at present Administrative staff's timing inspection or reports, can not accomplish (YUN weekly assembly etc. of deploying to ensure effective monitoring and control of illegal activities in all directions, in time, and one kind is built based on background The city of mould is set up a stall detection method, patent of invention, CN108012117A).In recent years, country make energetically smart city, Its net engineering, so that monitoring is throughout each corner in city, but only a small amount of monitoring is set up a stall supervision for city, and city is set up a stall It sets up an office and manages the main mode for using personnel's On-line monitor, this results in supervision not in time, not comprehensively, and then leads to human resources Waste and monitoring device utilize not in place.This assault formula, campaign-styled management simultaneously are easy to cause government administration section and society The conflict of the common people seriously affects the building of harmonious society.
The present invention uses video geography fence technology, proposes a kind of intellectualized detection manager towards phenomenon violating the regulations of setting up a stall Method solves to ask this field way to manage is not perfect, technological means is immature etc. at present to change traditional cities management mode Topic.
Video geography fence (Geo-fencing) is to be referred to based on a kind of new opplication of location-based service (LBS) development with void Quasi- fence crosses a virtual geographical frontier, when target entrance or leaves some specific geographical area, or in the area When movable, video sensor terminal can receive corresponding information on services, notice and warning automatically, and (Huang Jun is based on Android system Design and realization [D] University Of Ningbo of the long-range monitoring of system with control system, 2014.).Video geography fence technology has double To the features such as linkage, Passive Positioning, target be visual, the coordinates of targets of video geography fence border and acquisition can be realized in video Linkage under spatial scene shows and stores;This passive type positioning is so that the service of fence or monitoring objective are more comprehensive simultaneously Change and generalization, will not generate by sensor hardware equipment and target itself behavior institute and caused by fence target omit, And including the positioning visual information abundant such as target geographic coordinate, foreground information, color characteristic, geometric profile, for city Management etc. has significant application value.
Currently, video geography fence technology can be widely applied to multiple fields, and such as in safety-security area, video geography fence Monitoring objective be mainly pedestrian, vehicle and important item, it is intended to protect public property safety and citizen's personal safety;In traffic In field, the monitoring objective of video geography fence is motor vehicles, is mainly carried out to the unlawful practice on traffic route pre- Alert and timely progress information push.In field of power system, video geography fence by being with power line three-dimensional buffer Fence boundary position, when object enter the buffer area and be touching power line quick response, alarm in advance, support personnel safety (Wang Sining video geography fence and application [D] Nanjing Normal University, 2015.).
Summary of the invention
Critical issue to be solved by this invention is the deficiency for traditional cities management mode, by the way that geography fence is arranged And sdi video is mapped to, facility is managed to the flowing in video and is identified and is positioned, judges it and the ground established The positional relationship between fence is managed, violating the regulations set up a stall is determine whether according to fence rule.Therefore, the invention proposes one kind to be based on The detection method violating the regulations of setting up a stall of video geography fence.
The invention discloses a kind of based on video geography fence it is violating the regulations set up a stall monitoring method the following steps are included:
Step 1: according to the mapping relations in video image space and geographical space, by pre-set geography fence region Video image space is mapped to, the two-way linkage of video image space and geography fence region is established;
Step 2: the positional relationship of facility's objectives and geography fence region is managed according to the flowing in monitoring video flow, is sentenced It is disconnected to whether there is violation phenomenon.
Further, geography fence region is determined according to the fundamental of geography fence, the fundamental includes monitoring Time, monitoring place, monitoring objective and Monitoring Rules.
Further, the obtaining step of the mapping relations of the video image space and geographical space includes:
Obtain monitoring video flow in any piece image, on this image with the two-dimensional map of pre-rendered monitoring area Upper acquisition corresponding dot pair;
According to corresponding dot pair on image and two-dimensional map corresponding image coordinate and geographical coordinate, establish video image sky Between and geographical space mapping relations, obtain corresponding homography matrix;
Described the step of pre-set geography fence region is mapped to video image space includes: according to homography matrix Inverse matrix, geography fence region is mapped into video image space.
Further, the quantity of the corresponding dot pair is no less than 4 pairs, and corresponding dot pair cannot be all conllinear.
Further, the place of the geography fence region and rule pass through the typing of element text information and restricted language solution Analysis method, which combines, to be determined.
Further, the identification step of the flowing operation facility's objectives in the monitoring video flow includes:
Facility's objectives picture is managed according to flowing, flowing is established and manages facility's objectives image data set, and target is carried out Mark;
Using SSD model and TensorFlow frame, facility's objectives image data set is managed to flowing and is trained, is obtained Corresponding training result model;
According to training result model, each frame image in monitoring video flow is detected, determines flowing warp therein Facility's objectives are sought, and obtain location information.
The utility model has the advantages that the characteristics of the present invention is based on video geography fences, introducing the technology in city management field can be right The violating the regulations phenomenon of setting up a stall in city carries out more scientific, effectively and reasonably monitoring management.Meanwhile it is violating the regulations based on video geography fence Monitoring method of setting up a stall computational efficiency is high, notice setting fence fundamental, being capable of automatic discrimination and bullet when violation phenomenon occurs Warning window out improves the high efficiency, real-time and accuracy of monitoring.
Detailed description of the invention
Fig. 1 is geography fence element regulation schematic diagram of the present invention;
Fig. 2 is breakfast cart data set mark schematic diagram of the present invention;
Fig. 3 is general effect schematic diagram of the present invention.
Specific embodiment
The present invention is further explained with reference to the accompanying drawings and examples.
Basic ideas of the invention: the large-scale map near monitoring crossing is obtained;Monitoring video flow is accessed, it is obtained In piece image, acquire corresponding dot pair on video image and two-dimensional map, obtain the image coordinate of same place and geographical sit Mark, it is established that the mapping relations in video image space and geographical space;Fence place element, target type element, time are set Element and fence rule;Using homography matrix, the fence corresponding position that setting is completed is mapped into sdi video, establishes view The two-way linkage in frequency space and geography fence;By deep learning target identification method, to the flowing in video manage facility into Row identifies and determines its specific location;Judge that its positional relationship between established geography fence determines it if breaking the rules It sets up a stall to be violating the regulations.
Embodiment 1:
The pendulum violating the regulations based on video geography fence of the present embodiment is set up a stall monitoring method, comprising the following steps:
Step 1: drawing the ground that the scale bar near video monitoring regional is 1:200 using topographic map drawing method Figure.
Second step, the foundation in video image space and geographical space mapping relations: the mapping relations are established is first Monitoring video flow is accessed, piece image therein is obtained, corresponding dot pair is acquired on video image and two-dimensional map, is obtained of the same name The image coordinate and geographical coordinate of point, it is established that the mapping relations in video image space and geographical space obtain corresponding list and answer Matrix.Wherein, corresponding dot pair quantity should be no less than four pairs, and cannot be all conllinear.
Above-mentioned homography matrix calculation method are as follows:
The image coordinate and geographical coordinate of same place are obtained, is indicated are as follows:
Wherein, s is proportionality coefficient, and H is homography matrix, it includes the physical conversion and camera intrinsic parameter square between image Battle array two parts, obtain following linear equation:
Corresponding dot pair is substituted into above-mentioned equation, homography matrix H can be obtained.
Third step, the extraction of fence fundamental: the fence fundamental includes monitoring place, monitoring time, prison Target, Monitoring Rules etc. are surveyed, the method being combined with each other by the modes such as the typing of element text information and restricted language parsing is established The information such as region, the rule of geography fence, fence area determine its boundary position, fence effect by the way of hand drawn Time and specific rules are determined by drop-down option, and initial time " 7:00 " such as is arranged, and are terminated time " 9:00 ", fence rule For " forbidding intersecting with face ", for indicating the region no peddler.
The two-way linkage of 4th step, sdi video and geography fence: the geography fence location information that will be obtained in third step, By obtaining the inverse matrix of homography matrix in second step, geography fence is mapped into sdi video, establishes sdi video and geography The two-way linkage of fence.
The identification of facility is managed in 5th step, flowing:
Acquisition flowing manages the foundation flowing of facility's objectives picture and manages facility's objectives image data set, utilizes label_img Tool is labeled interesting target;
Using SSD (Single-Shot Detector) model and TensorFlow frame, to the data set of foundation into Row training, obtains corresponding training result model;
Using training result model, each frame image in monitoring video flow is detected, determines flowing booth therein Position target, and obtain its position;
6th step, the judgement of behavior violating the regulations of setting up a stall.
The behavior decision rule violating the regulations of setting up a stall is to judge pitch flow according to the positional relationship between point and polygon It is determined as with the positional relationship of geography fence determined in second step if target is located in fence area in set period of time Violation phenomenon;Conversely, being not belonging to violating the regulations set up a stall.
Embodiment 2:
The first step, relevant device prepare: one portable notebook computer of preparation, and high definition monitoring camera one.
The acquisition of second step, large-scale map draws the ratio near video monitoring regional using topographic map drawing method The map that example ruler is 1:200.
Third step, the method acquired by same place, obtain the respective coordinates point pair of video image and geographical space, calculate The homography matrix of video image and map space constructs the mutual mapping relations of video image and geographical space.
4th step determines fence element, and the position of fence boundary is determined by the way of the choosing of manual frame, is selected by drop-down Item determines fence action time and specific rules, as shown in Figure 1, realizing the elements recognition to geography fence.
5th step, using the inverse matrix for acquiring homography matrix in third step, the geographical coordinate of fence boundary is converted into prison Image coordinate is controlled, and draws fence area, thus establishes the two-way linkage of sdi video and geography fence.
6th step identifies pitch flow using deep learning method, and the present embodiment is by taking breakfast cart as an example:
(1) breakfast cart image data set is initially set up, from network collection correlation breakfast cart image, data volume sample size is about 10000, and target is labeled using label_img tool, as shown in Figure 2.
(2) SSD model and TensorFlow frame are used, the data set established in (1) is utilized to be trained, training About 10000 step of process, obtains corresponding training result model.
(3) the training result model in (2) is utilized, each frame image in video flowing is detected, is determined therein Pitch flow target, and obtain its image coordinate.
7th step, according to the positional relationship between point and polygon, judge to determine geography in pitch flow and second step The positional relationship of fence is set up a stall if it within fence, determines that it is violating the regulations.

Claims (6)

1. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence, it is characterised in that: the following steps are included:
Step 1: according to the mapping relations in video image space and geographical space, pre-set geography fence region is mapped To video image space, the two-way linkage of video image space and geography fence region is established;
Step 2: managing the positional relationship of facility's objectives and geography fence region according to the flowing in monitoring video flow, and judgement is It is no that there are violation phenomenons.
2. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence according to claim 1, it is characterised in that: root Geography fence region is determined according to the fundamental of geography fence, and the fundamental includes monitoring time, monitoring place, monitoring mesh Mark and Monitoring Rules.
3. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence according to claim 1, it is characterised in that: institute The obtaining step for stating the mapping relations of video image space and geographical space includes:
Any piece image in monitoring video flow is obtained, is adopted on this image on the two-dimensional map of pre-rendered monitoring area Collect corresponding dot pair;
According to corresponding dot pair on image and two-dimensional map corresponding image coordinate and geographical coordinate, establish video image space and The mapping relations of geographical space obtain corresponding homography matrix;
Described the step of pre-set geography fence region is mapped to video image space includes: according to the inverse of homography matrix Geography fence region is mapped to video image space by matrix.
4. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence according to claim 3, it is characterised in that: institute The quantity for stating corresponding dot pair is no less than 4 pairs, and corresponding dot pair cannot be all conllinear.
5. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence according to claim 2, it is characterised in that: institute The place and rule for stating geography fence region are combined by the typing of element text information and restricted language analytic method to be determined.
6. a kind of monitoring method violating the regulations of setting up a stall based on video geography fence according to claim 1, it is characterised in that: institute State in monitoring video flow flowing manage facility's objectives identification step include:
Facility's objectives picture is managed according to flowing, flowing is established and manages facility's objectives image data set, and target is labeled;
Using SSD model and TensorFlow frame, facility's objectives image data set is managed to flowing and is trained, obtained corresponding Training result model;
According to training result model, each frame image in monitoring video flow is detected, determines that flowing operation therein is set Target is applied, and obtains location information.
CN201910350804.7A 2019-04-28 2019-04-28 A kind of monitoring method violating the regulations of setting up a stall based on video geography fence Pending CN110166744A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910350804.7A CN110166744A (en) 2019-04-28 2019-04-28 A kind of monitoring method violating the regulations of setting up a stall based on video geography fence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910350804.7A CN110166744A (en) 2019-04-28 2019-04-28 A kind of monitoring method violating the regulations of setting up a stall based on video geography fence

Publications (1)

Publication Number Publication Date
CN110166744A true CN110166744A (en) 2019-08-23

Family

ID=67638816

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910350804.7A Pending CN110166744A (en) 2019-04-28 2019-04-28 A kind of monitoring method violating the regulations of setting up a stall based on video geography fence

Country Status (1)

Country Link
CN (1) CN110166744A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111080952A (en) * 2020-01-13 2020-04-28 贵州安防工程技术研究中心有限公司 Feedback type regional intrusion detection method and system based on real-time video and rules
CN112153107A (en) * 2020-08-12 2020-12-29 上海新爱季信息技术有限公司 Stall management method
CN112288613A (en) * 2020-11-14 2021-01-29 浙江信电技术股份有限公司 City management comprehensive law enforcement system
CN113676696A (en) * 2020-05-14 2021-11-19 杭州萤石软件有限公司 Target area monitoring method and system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090315712A1 (en) * 2006-06-30 2009-12-24 Ultrawave Design Holding B.V. Surveillance method and system using object based rule checking
CN108012117A (en) * 2017-11-30 2018-05-08 江西洪都航空工业集团有限责任公司 A kind of city based on background modeling is set up a stall detection method
CN105005772B (en) * 2015-07-20 2018-06-12 北京大学 A kind of video scene detection method
CN108304787A (en) * 2018-01-17 2018-07-20 河南工业大学 Road target detection method based on convolutional neural networks
CN109003439A (en) * 2018-08-30 2018-12-14 新华三技术有限公司 A kind of peccancy detection method and device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090315712A1 (en) * 2006-06-30 2009-12-24 Ultrawave Design Holding B.V. Surveillance method and system using object based rule checking
CN105005772B (en) * 2015-07-20 2018-06-12 北京大学 A kind of video scene detection method
CN108012117A (en) * 2017-11-30 2018-05-08 江西洪都航空工业集团有限责任公司 A kind of city based on background modeling is set up a stall detection method
CN108304787A (en) * 2018-01-17 2018-07-20 河南工业大学 Road target detection method based on convolutional neural networks
CN109003439A (en) * 2018-08-30 2018-12-14 新华三技术有限公司 A kind of peccancy detection method and device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王思宁: "视频地理围栏与应用", 《南京师范大学》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111080952A (en) * 2020-01-13 2020-04-28 贵州安防工程技术研究中心有限公司 Feedback type regional intrusion detection method and system based on real-time video and rules
CN113676696A (en) * 2020-05-14 2021-11-19 杭州萤石软件有限公司 Target area monitoring method and system
CN112153107A (en) * 2020-08-12 2020-12-29 上海新爱季信息技术有限公司 Stall management method
CN112288613A (en) * 2020-11-14 2021-01-29 浙江信电技术股份有限公司 City management comprehensive law enforcement system

Similar Documents

Publication Publication Date Title
Xia et al. Analyzing spatial relationships between urban land use intensity and urban vitality at street block level: A case study of five Chinese megacities
CN113065000B (en) Multisource heterogeneous data fusion method based on geographic entity
CN110166744A (en) A kind of monitoring method violating the regulations of setting up a stall based on video geography fence
Wang et al. Machine learning-based regional scale intelligent modeling of building information for natural hazard risk management
CN100533486C (en) Digital city full-automatic generating method
CN112418674A (en) City multi-source data-based street space quality measure evaluation method and system
Wu et al. Using street view images to examine the association between human perceptions of locale and urban vitality in Shenzhen, China
CN116227834A (en) Intelligent scenic spot digital platform based on three-dimensional point cloud model
KR101937940B1 (en) Method of deciding cpted cctv position by big data
CN106991218A (en) A kind of visualization system emulated based on CBC
Song et al. Real-time monitoring for crowd counting using video surveillance and GIS
CN108304536A (en) A kind of geographical environmental simulation of the geographical environmental element of coupling and predicting platform
Qiu et al. Disaggregating population data for assessing progress of SDGs: methods and applications
CN114723904A (en) Method, system, computer device and storage medium for dynamic management of airport data
Chen et al. Earthquake-induced building damage mapping based on multi-task deep learning framework
Huang et al. Comprehensive urban space representation with varying numbers of street-level images
CN116994156B (en) Landslide hidden danger comprehensive remote sensing identification method, system, equipment and medium
Hu et al. Urban landscape information atlas and model system based on remote sensing images
Li et al. Evaluation of tree shade effectiveness and its renewal strategy in typical historic districts: A case study in Harbin, China
Yan et al. Evaluating simulated visible greenness in urban landscapes: An examination of a midsize US city
Xu et al. Analysis of the spatiotemporal expansion and pattern evolution of urban areas in Anhui Province, China, based on nighttime light data
CN112347926B (en) High-resolution image city village detection method based on building morphology distribution
Wu Intelligent city evaluation system
Dou et al. Construction of smart community based on GIS and tilt photogrammetry
Zhang et al. Video surveillance GIS: A novel application

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190823