CN106446807A - Well lid theft detection method - Google Patents

Well lid theft detection method Download PDF

Info

Publication number
CN106446807A
CN106446807A CN201610812920.2A CN201610812920A CN106446807A CN 106446807 A CN106446807 A CN 106446807A CN 201610812920 A CN201610812920 A CN 201610812920A CN 106446807 A CN106446807 A CN 106446807A
Authority
CN
China
Prior art keywords
well lid
value
image
video sequence
detection method
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
CN201610812920.2A
Other languages
Chinese (zh)
Inventor
邹泽东
龙学军
韩明燕
毛河
周剑
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.)
Chengdu Tongjia Youbo Technology Co Ltd
Original Assignee
Chengdu Tongjia Youbo Technology Co Ltd
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 Chengdu Tongjia Youbo Technology Co Ltd filed Critical Chengdu Tongjia Youbo Technology Co Ltd
Priority to CN201610812920.2A priority Critical patent/CN106446807A/en
Publication of CN106446807A publication Critical patent/CN106446807A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/40Scenes; Scene-specific elements in video content
    • G06V20/41Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/24Aligning, centring, orientation detection or correction of the image
    • G06V10/245Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/40Scenes; Scene-specific elements in video content
    • G06V20/44Event detection

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Closed-Circuit Television Systems (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a well lid theft detection method. The method comprises the steps of obtaining a video sequence image containing a well lid from a monitoring device; determining an ellipse coinciding with the edge of the well lid through an image processing algorithm; performing calculation according to features in an ellipse range of the video sequence image to obtain a reference feature value; obtaining the video sequence image from the monitoring device at an interval of first preset time, and performing calculation according to the features in the ellipse range of the video sequence image to obtain a comparison feature value; performing comparison to obtain a difference value between the comparison feature value and the reference feature value, if the difference value is greater than a reference error value, adding 1 to an accumulator, and if the difference value is smaller than the reference error value, resetting the accumulator; and if a count of the accumulator exceeds a reference threshold, outputting a theft signal. According to the method, a detection algorithm provided by the invention still can perform effective identification and locating on a deformed well lid image, the detection speed and accuracy is increased and improved, the calculation complexity is lowered, and a calculation mechanism of the accumulator can effectively avoid a false detection phenomenon.

Description

A kind of stolen detection method of well lid
Technical field
The present invention relates to a kind of stolen detection method of well lid.
Background technology
Well lid facility is the critical asset of urban construction, and city management is significant.But well lid is lost and is asked Topic is effectively solved always, it will produce great potential safety hazard to pedestrian and vehicle.Therefore it is real-time that a kind of energy is studied The technology that ground is monitored to well lid state and loses alarm in well lid has very important significance.
The solution traditional to this problem has two kinds, a kind of be by improve well lid manufacturing process, well lid is made Antitheft type, is unfavorable for stealing, or directly welds together well lid with road surface.Another kind is to increase on well lid to report to the police Sensing device, has the situation of illegal unlatching when well lid is detected, and alarm signal will be transferred to relevant staff by the device. Although both the above mode can improve the stolen situation of well lid to a certain extent, however it is necessary that put into substantial amounts of manpower and wealth Power, and further increase the stolen cost of well lid.
Content of the invention
In view of this, it is an object of the present invention to provide a kind of judge the stolen well lid detection method of well lid by graphical analyses.
In order to solve above-mentioned technical problem, the technical scheme is that:
A kind of stolen detection method of well lid,
Step 1, obtains the video sequence image comprising well lid from monitoring device;
Step 2, determines, by image processing algorithm, the elliptic region for overlapping with well lid edge;
Step 3, the image feature value described in the video sequence image in calculation procedure 1 in elliptic region obtains reference characteristic Value;
Step 4, the first Preset Time of interval obtains video sequence image from monitoring device, and calculates in the video sequence image Image feature value in elliptic region is to obtain contrast characteristic's value;
Step 5, calculates the difference of contrast characteristic's value and the reference characteristic value, if difference is cumulative more than during fiducial error value The count value of device adds one, if difference is reset less than the count value of accumulator during fiducial error value;
Stolen signal is exported if the count value of accumulator exceedes baseline threshold;If the count value of accumulator is less than baseline threshold Return to step 4.
So arrange, first, be analyzed by the image for having monitoring device in the presence of well lid to be returned, and determine well lid Position, then an ellipse figure that can represent well lid is obtained by image processing algorithm(Restriction well due to shooting angle Lid is typically all ellipse in the picture), then by way of a kind of eigenvalue calculation, eigenvalue in the occluded ellipse is calculated, Using this eigenvalue as the eigenvalue of standard, detect, for successive image, the use that compares, and in subsequent diagram, as long as at this Figure in individual ellipse, all carries out eigenvalue calculation, finally obtains the eigenvalue of a contrast, so, it is possible to two Individual eigenvalue asks poor, so it may determine that in image, well lid whether there is, but in practice, can be because well lid be blocked Situation makes eigenvalue change to be caused to judge well lid now not in the picture, but can not be determined directly as stolen, so this Invention is also provided with accumulator, when the data of multiple passback cause accumulative frequency to reach threshold value, is judged as that well lid is stolen, so Judge and image procossing, a preferably effect can be obtained.
Further:Step 2 includes
Step 2-1, carries out first differential to video sequence image and obtains gradient image;
Step 2-2, determines all edge pattern in the gradient image to the gradient image by edge detection algorithm;
Step 2-3, filters out the ellipse figure for being shaped as ellipse in all edge pattern;
Step 2-4, determines area closest to the ellipse figure of area reference value and with the ellipse figure as elliptic region.
For carrying out first differential to video sequence image for discrete image, the mathematic(al) representation of first differential Difference quite with two neighbors, so processes and the marginalisation of image is become apparent from, conveniently carry out rim detection calculation Method determines edge pattern, and thus can judge whether all figures are ellipse figure according to edge pattern, and can basis The position of monitoring device and angle, in conjunction with the area of actual well lid, design the area reference value of an estimation, after process When figure is closest to the area reference value, the ellipse figure that the position at well lid place is formed is judged as.
Further:In step 2-1, first differential is carried out to video sequence image by Sobel operator.According to selected Gradient operator difference, effect may be different, but ultimate principle will not change.Common operator have Sobel, Roberts, The operators such as Prewitt.Here image is processed using Sobel operator.The result of differential process is overlapped so that image border Display becomes apparent from.
Further:In step 2-2, all edge pattern in gradient image are determined by Canny edge detection algorithm.Logical Crossing Canny rim detection carries out the process to image, and image is more smoothed, it is adaptable to determine ellipse figure, and data processing amount Less.
Further:In step 2-2, after determining all edge pattern, screening excludes edge of the area less than area threshold Figure.Area threshold is less than area reference value, and the installation site also according to monitoring device and the actual size of well lid determine, And area threshold could be arranged to the 0.5 of area benchmark, it is ensured that well lid will not be excluded.
Further:Step 2-3 includes
Step 2-3-1, the edge pattern is divided into some curved lines, is mutual with two curved lines for having common continuity point Continuous curved line;
Step 2-3-2, calculates the anglec of rotation of each curved line respectively, and all curved lines are carried out with feature group division:If Mutually continuous curved line is identical in the anglec of rotation of the corresponding continuity point, then the two curved lines are judged as same spy Levy group;
Step 2-3-3, the curved line of same feature group is obtained ellipse figure as ellipse fitting.
As ellipse has a more obvious feature, the rotational angle on per section of camber line be continuous, and special by this Levy, it is possible to which other non-oval figures are carried out screening exclusion, obtain multigroup qualified packet and be assured that one is ellipse Circle diagram shape, relatively simple facility, and be fitted then and oval home position can be determined by a plurality of curved line, realize oval position Determination.
Further:In step 2-3-2, if mutually continuous curved line rotational angle is discontinuous, its continuity point is broken Open.In step 2-3-2, screening excludes curved line of the arc length less than arc length threshold value.So this algorithm, can one step ahead will arc The discontinuous figure of shape line rotational angle is excluded, while single curved line can also be formed, the less curved line of arc length is entered Row is excluded, and reduces the complexity that output is processed.
Further:The feature calculation of contrast characteristic's value and reference characteristic value is HOG feature calculation.Calculated by HOG, make Eigenvalue obvious difference is obtained, is easy to judge.
Further:In step 1, gaussian filtering process is adopted to the video sequence image for gathering;And/or in step 4, right The video sequence image of collection adopts gaussian filtering process.Can eliminate and cause in a large number, remove noise jamming.
The technology of the present invention effect major embodiment is in the following areas:Well lid detection part mainly using the marginal information of image, Can avoid shooting the impact for causing image under different illumination conditions.The angle of shooting is different from place, can cause image In well lid there is certain deformation, detection algorithm presented here can still provide for effective knowledge to the well lid image for deforming upon Not with positioning, and improve speed and the accuracy rate of detection.
Well lid status monitoring part optimizes correlation technique, reduces computation complexity, realizes monitoring well lid shape in real time State, the computing mechanism of accumulator can be prevented effectively from miss detection, be significant to putting in real world applications.
Description of the drawings
Fig. 1:Flow chart of the present invention.
Specific embodiment
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail, so that technical solution of the present invention is more Should be readily appreciated that and grasp.
With reference to shown in Fig. 1, a kind of stolen detection method of well lid,
Step 1, obtains the video sequence image comprising well lid from monitoring device;Gauss is adopted to the video sequence image for gathering Filtering Processing.Video acquisition is by photographic head Real-time Collection video image, and video sequence image is passed to the detection of terminal well lid Module., as the environment residing for monitor video is complicated and changeable, there is substantial amounts of interference noise, equipment itself in Image semantic classification Noise may be brought to image, therefore first have to adopt image gaussian filtering, remove noise jamming.
Step 2, determines an ellipse for overlapping with well lid edge by image processing algorithm;
Including step 2-1, first differential is carried out to video sequence image and obtains gradient image;Especially by Sobel operator to regarding Frequency sequence image carries out first differential.Grad enhancement, is the marginal information for strengthening image, respectively the x and y direction of picture is respectively entered Row first differential, for discrete image, the mathematical expression of first differential equivalent to the difference of two neighbors, according to The gradient operator difference of selection, effect may be different, but ultimate principle will not change.Common operator have Sobel, The operators such as Roberts, Prewitt.Here image is processed using Sobel operator.The result of differential process is overlapped so that Image border shows and becomes apparent from.
Step 2-2, determines all edge pattern in gradient image to gradient image by edge detection algorithm;Especially by Canny edge detection algorithm determines all edge pattern in gradient image, and after determining all edge pattern, it is little that screening excludes area Edge pattern in area threshold.
Step 2-3, filters out the ellipse figure in all edge pattern;Including
Step 2-3-1, filters out some curved lines from edge pattern;
Step 2-3-2, according to the anglec of rotation for calculating each curved line respectively, if mutually continuous curved line rotational angle is continuous Then it is divided into same group;If mutually continuous curved line rotational angle is discontinuous, its continuity point is disconnected, while screen excluding arc length Curved line less than arc length threshold value.
Step 2-3-3, will obtain ellipse figure with the curved line of group as ellipse fitting.
Step 2-4, determines area closest to the ellipse figure of area reference value.
Well lid is detected
First in step 1 acquired results, primary screening is carried out according to size, exclude the too small region of area, remaining wheel Exterior feature also not all meets the geometry eigenellipse structure of well lid, realizes the detection to well lid image herein, concrete step Rapid as follows:A. it is the edge of elliptical shape to exclude first, calculate the rotation of arc between two points of certain intervals on edge Turn direction, two adjacent groups are compared, if the anglec of rotation is different, the line between this two sections of arcs is disconnected.B. actual feelings are combined Condition, the arc of length very little in the result of A is removed.C. midpoint and the arc of the arc for remaining, after step B, are extracted Two end points lines midpoint, calculate the angle of inclination of arc according to the two points, in conjunction with the corresponding home position of arc, will Oval arc may be constituted and be divided into one group, then to making ellipse fitting per group.Well in image is gone out according to ellipse area size detection Lid position.
Step 3, obtains reference characteristic value according to the feature calculation in oval scope in the video sequence image in step 1;
Step 4, the first Preset Time of interval obtains video sequence image from monitoring device, and according in the video sequence image Feature calculation in oval scope obtains contrast characteristic's value;The feature calculation of contrast characteristic's value and reference characteristic value is HOG feature Calculate.Gaussian filtering process is adopted to the video sequence image for gathering.The setting of the first Preset Time can be according to algorithm reality Demand is arranged, and could be arranged to 1 minute, it is also possible to be set to 1 hour, while can be according to the different benchmark of time different acquisitions Eigenvalue, for example, obtain every a hour and update a reference characteristic value, and detection per minute is once.
Step 5, compares the difference for obtaining contrast characteristic's value and reference characteristic value, if difference is cumulative more than during fiducial error value Device adds one, if difference is reset less than accumulator during fiducial error value;The design of Fiducial Value of Error for Power, then relevant with feature value-based algorithm, Experiment method can be passed through obtained according to practical situation, it is also possible to by calculating relatively reasonable fiducial error under the scene Value.
Stolen signal is exported if accumulator count exceedes baseline threshold;Return if accumulator count is less than baseline threshold Step 4.Status monitoring, after the position for obtaining well lid, monitors the well lid state of the position, judges whether to lose.A. feature Extract, the HOG feature of the oval scope for obtaining in calculation procedure 3 on former video sequence image is used as the mark sheet of the well lid Show.HOG feature full name is histograms of oriented gradients feature, and it is the gradient direction of the regional area by calculating and statistical picture Rectangular histogram carrys out constitutive characteristic.B. after feature compares intervals, calculate the HOG feature of the range areas again, and with step Result of calculation in rapid 3 is compared.If result difference is more than setting value, accumulator adds 1, if difference very little, accumulator resets. C. the step of well lid condition adjudgement repeats certain number of times 4, if the value of accumulator exceedes given threshold, judges that well lid may be lost Lose, alarm signal is sent, otherwise returns to step 4.Baseline threshold to be designed at detection time relevant, can arrange according to demand, It is exactly to compare threshold value in a hour and baseline threshold difference is larger as the baseline threshold of detection in a minute is set to 60 When more than a hour, then be judged as stolen, subsequently stolen signal can be sent by way of wired or wireless to appoint Meaning terminal is reported to the police.
Certainly, the representative instance of the above simply present invention, in addition, the present invention can also have other multiple to be embodied as The technical scheme that mode, all employing equivalents or equivalent transformation are formed, all falls within the scope of protection of present invention.

Claims (10)

1. the stolen detection method of a kind of well lid, it is characterised in that:
Step 1, obtains the video sequence image comprising well lid from monitoring device;
Step 2, determines, by image processing algorithm, the elliptic region for overlapping with well lid edge;
Step 3, the image feature value described in the video sequence image in calculation procedure 1 in elliptic region obtains reference characteristic Value;
Step 4, the first Preset Time of interval obtains video sequence image from monitoring device, and calculates in the video sequence image Image feature value in elliptic region is to obtain contrast characteristic's value;
Step 5, calculates the difference of contrast characteristic's value and the reference characteristic value, if difference is cumulative more than during fiducial error value The count value of device adds one, if difference is reset less than the count value of accumulator during fiducial error value;
Stolen signal is exported if the count value of accumulator exceedes baseline threshold;If the count value of accumulator is less than baseline threshold Return to step 4.
2. the stolen detection method of a kind of well lid as claimed in claim 1, it is characterised in that:
Step 2 includes
Step 2-1, carries out first differential to video sequence image and obtains gradient image;
Step 2-2, determines all edge pattern in the gradient image to the gradient image by edge detection algorithm;
Step 2-3, filters out the ellipse figure for being shaped as ellipse in all edge pattern;
Step 2-4, determines area closest to the ellipse figure of area reference value and with the ellipse figure as elliptic region.
3. the stolen detection method of a kind of well lid as claimed in claim 2, it is characterised in that:
In step 2-1, first differential is carried out to video sequence image by Sobel operator.
4. the stolen detection method of a kind of well lid as claimed in claim 2, it is characterised in that:
In step 2-2, all edge pattern in the gradient image are determined by Canny edge detection algorithm.
5. the stolen detection method of a kind of well lid as claimed in claim 2, it is characterised in that:
In step 2-2, after determining all edge pattern, screening excludes edge pattern of the area less than area threshold.
6. the stolen detection method of a kind of well lid as claimed in claim 2, it is characterised in that:
Step 2-3 includes
Step 2-3-1, the edge pattern is divided into some curved lines, is mutual with two curved lines for having common continuity point Continuous curved line;
Step 2-3-2, calculates the anglec of rotation of each curved line respectively, and all curved lines are carried out with feature group division:If Mutually continuous curved line is identical in the anglec of rotation of the corresponding continuity point, then the two curved lines are judged as same spy Levy group;
Step 2-3-3, the curved line of same feature group is obtained ellipse figure as ellipse fitting.
7. the stolen detection method of a kind of well lid as claimed in claim 6, it is characterised in that:
In step 2-3-2, if mutually continuous curved line rotational angle is discontinuous, its continuity point is disconnected.
8. the stolen detection method of a kind of well lid as claimed in claim 1, it is characterised in that:In step 2-3-2, screening excludes arc The long curved line less than arc length threshold value.
9. the stolen detection method of a kind of well lid as claimed in claim 1, it is characterised in that:Contrast characteristic's value and reference characteristic value Computational methods be HOG feature calculation.
10. the stolen detection method of a kind of well lid as claimed in claim 1, it is characterised in that:In step 1, to the video for gathering Sequence image adopts gaussian filtering process;And/or in step 4, gaussian filtering process is adopted to the video sequence image for gathering.
CN201610812920.2A 2016-09-09 2016-09-09 Well lid theft detection method Pending CN106446807A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610812920.2A CN106446807A (en) 2016-09-09 2016-09-09 Well lid theft detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610812920.2A CN106446807A (en) 2016-09-09 2016-09-09 Well lid theft detection method

Publications (1)

Publication Number Publication Date
CN106446807A true CN106446807A (en) 2017-02-22

Family

ID=58167521

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610812920.2A Pending CN106446807A (en) 2016-09-09 2016-09-09 Well lid theft detection method

Country Status (1)

Country Link
CN (1) CN106446807A (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107256413A (en) * 2017-06-14 2017-10-17 广东工业大学 A kind of article monitoring method and device
CN108573234A (en) * 2018-04-19 2018-09-25 孙磊 City manhole cover miss status identifying system
CN108765829A (en) * 2018-07-06 2018-11-06 江西洪都航空工业集团有限责任公司 A kind of detection of municipal stolen articles and alarm method based on intelligent video analysis
CN109564632A (en) * 2017-05-24 2019-04-02 深圳配天智能技术研究院有限公司 A kind of visible detection method, equipment, system and the device with store function
CN110097092A (en) * 2019-03-04 2019-08-06 厦门攸信信息技术有限公司 A kind of location-authentication device and method
CN111339905A (en) * 2020-02-22 2020-06-26 郑州铁路职业技术学院 CIM well lid state visual detection system based on deep learning and multi-view angle
CN111402192A (en) * 2018-12-28 2020-07-10 杭州海康威视数字技术股份有限公司 Inspection well cover detection method and device
CN112668478A (en) * 2020-12-29 2021-04-16 河南橡树智能科技有限公司 Well lid monitoring method and device, electronic equipment and storage medium
CN113129564A (en) * 2019-12-30 2021-07-16 宇龙计算机通信科技(深圳)有限公司 Well lid safety monitoring method and device, storage medium and electronic equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101571982A (en) * 2009-05-11 2009-11-04 宁波海视智能系统有限公司 Method for judging stolen articles in video monitoring range
CN102254395A (en) * 2011-06-07 2011-11-23 北京引创科技有限责任公司 Article theft detecting and alarming device and method based on intelligent video analysis
CN202495068U (en) * 2012-03-23 2012-10-17 上海海事大学 Video surveillance image moving target detection device under foggy weather condition
US20130169787A1 (en) * 2011-12-28 2013-07-04 Hamamatsu Photonics K.K. Image processing device, imaging device, microscope device, image processing method, and image processing program
CN105894504A (en) * 2016-03-30 2016-08-24 上海海事大学 Manhole cover missing detection method based on images

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101571982A (en) * 2009-05-11 2009-11-04 宁波海视智能系统有限公司 Method for judging stolen articles in video monitoring range
CN102254395A (en) * 2011-06-07 2011-11-23 北京引创科技有限责任公司 Article theft detecting and alarming device and method based on intelligent video analysis
US20130169787A1 (en) * 2011-12-28 2013-07-04 Hamamatsu Photonics K.K. Image processing device, imaging device, microscope device, image processing method, and image processing program
CN202495068U (en) * 2012-03-23 2012-10-17 上海海事大学 Video surveillance image moving target detection device under foggy weather condition
CN105894504A (en) * 2016-03-30 2016-08-24 上海海事大学 Manhole cover missing detection method based on images

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
唐录洁: "雨天井盖缺失的计算机智能识别方法研究仿真", 《计算机仿真》 *
雷可: "一种基于图像的路面井盖破损自动检测方法", 《数字通信世界》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109564632A (en) * 2017-05-24 2019-04-02 深圳配天智能技术研究院有限公司 A kind of visible detection method, equipment, system and the device with store function
CN107256413A (en) * 2017-06-14 2017-10-17 广东工业大学 A kind of article monitoring method and device
CN108573234A (en) * 2018-04-19 2018-09-25 孙磊 City manhole cover miss status identifying system
CN108765829A (en) * 2018-07-06 2018-11-06 江西洪都航空工业集团有限责任公司 A kind of detection of municipal stolen articles and alarm method based on intelligent video analysis
CN111402192A (en) * 2018-12-28 2020-07-10 杭州海康威视数字技术股份有限公司 Inspection well cover detection method and device
CN111402192B (en) * 2018-12-28 2023-10-27 杭州海康威视数字技术股份有限公司 Inspection well cover detection method and inspection well cover detection device
CN110097092A (en) * 2019-03-04 2019-08-06 厦门攸信信息技术有限公司 A kind of location-authentication device and method
CN110097092B (en) * 2019-03-04 2021-06-18 厦门攸信信息技术有限公司 Position authentication device and method
CN113129564A (en) * 2019-12-30 2021-07-16 宇龙计算机通信科技(深圳)有限公司 Well lid safety monitoring method and device, storage medium and electronic equipment
CN111339905A (en) * 2020-02-22 2020-06-26 郑州铁路职业技术学院 CIM well lid state visual detection system based on deep learning and multi-view angle
CN111339905B (en) * 2020-02-22 2022-07-08 郑州铁路职业技术学院 CIM well lid state visual detection system based on deep learning and multiple visual angles
CN112668478A (en) * 2020-12-29 2021-04-16 河南橡树智能科技有限公司 Well lid monitoring method and device, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
CN106446807A (en) Well lid theft detection method
CN106652445B (en) A kind of road traffic accident method of discrimination and device
CN106652465B (en) Method and system for identifying abnormal driving behaviors on road
JP6571545B2 (en) Object detection apparatus and object detection method
CN105426864B (en) One kind being based on the matched Multi-lane Lines Detection method of equidistant marginal point
CN110738150B (en) Camera linkage snapshot method and device and computer storage medium
CN110718061B (en) Traffic intersection vehicle flow statistical method and device, storage medium and electronic equipment
US10699567B2 (en) Method of controlling a traffic surveillance system
KR19980079232A (en) Traffic monitoring device
CN105678288A (en) Target tracking method and device
CN112349087B (en) Visual data input method based on holographic perception of intersection information
Liu et al. Vision-based real-time lane marking detection and tracking
CN103116757B (en) A kind of three-dimensional information restores the road extracted and spills thing recognition methods
JP6226368B2 (en) Vehicle monitoring apparatus and vehicle monitoring method
CN105260701A (en) Front vehicle detection method applied to complex scene
CN106327880B (en) A kind of speed recognition methods and its system based on monitor video
CN111524350B (en) Method, system, terminal device and medium for detecting abnormal driving condition of vehicle and road cooperation
CN110889328A (en) Method, device, electronic equipment and storage medium for detecting road traffic condition
CN107516423B (en) Video-based vehicle driving direction detection method
CN106920253A (en) It is a kind of based on the multi-object tracking method for blocking layering
CN104267209B (en) Method and system for expressway video speed measurement based on virtual coils
CN112733598A (en) Vehicle law violation determination method and device, computer equipment and storage medium
Jung et al. An improved linear-parabolic model for lane following and curve detection
TWI730509B (en) Method of acquiring detection zone in image and method of determining zone usage
Lee et al. A cumulative distribution function of edge direction for road-lane detection

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into 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: 20170222