CN105554462A - Remnant detection method - Google Patents
Remnant detection method Download PDFInfo
- Publication number
- CN105554462A CN105554462A CN201510985628.6A CN201510985628A CN105554462A CN 105554462 A CN105554462 A CN 105554462A CN 201510985628 A CN201510985628 A CN 201510985628A CN 105554462 A CN105554462 A CN 105554462A
- Authority
- CN
- China
- Prior art keywords
- target
- remnant
- image
- detection method
- suspicious
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/2224—Studio circuitry; Studio devices; Studio equipment related to virtual studio applications
- H04N5/2226—Determination of depth image, e.g. for foreground/background separation
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a remnant detection method. The remnant detection method is characterized by comprising the following steps: (1), obtaining an image acquired at the front end according to an image photographed by a camera; (2), extracting foreground targets of the obtained image; (3), extracting suspicious remnant targets according to the characteristics of the foreground targets; (4), performing characteristic analysis of the extracted suspicious targets, and removing non-remnant targets; and (5), performing further characteristic analysis of the screened targets, and determining the remnant targets. The remnant detection method disclosed by the invention has the benefits that the problems of being slow in detection speed, high in complexity and the like due to a double-background model can be avoided; the remnant detection method can deal with a certain degree of target occlusion; therefore, leak detection of remnants can be reduced; false detection due to the problems, such as shadow and illumination, can be reduced; false detection due to the fact that an object in a scene moves can be eliminated; the problem that the false report and leak report rate is high in detection of the remnants can be solved; and the remnant detection accuracy is improved.
Description
(1) technical field
The invention belongs to video monitoring, Computer Vision and analysis and technical field of machine vision, particularly a kind of remnant object detection method.
(2) background technology
The situation is tense for world today's anti-terrorism, in order to ensure social public security, the generation stoping the public place attacks of terrorism such as bank, airport, subway, exhibition center, gymnasium, railway station, market has been extremely urgent task, and the main mode with parcel bomb of the attack of terrorism in these places occurs more, thus to these places to leave over that parcel detects be the indispensable function of intelligent video monitoring system.
Existing remnant object detection method mainly contains two kinds: the method that based target is followed the tracks of and the method that based target detects.
The method that based target is followed the tracks of: first detect foreground target, carry out real-time tracking to each target entered in video scene, according to information such as the movement locus of target and space-time characterisations, adopts corresponding algorithm to detect legacy.This method needs solution to enter, follow the tracks of, split, leave, the problem such as to block, and is not suitable for the occasion that the stream of people such as airport, station hall is crowded, scene is complicated; Cannot follow the tracks of by accurate marker when target is intersected and blocked; In addition, these class methods length consuming time, processing speed is low, is not suitable for real-time legacy and detects.
The method that based target detects: according to the moving target detected, utilize the statistical informations such as target travel attribute and space-time characterisation, set corresponding Time and place threshold value and differentiate.These class methods compare the method based on following the tracks of, without the need to following the tracks of all targets in scene, consuming time few, complexity is low, if use more suitable statistical information to carry out process after target being detected should be able to reach reasonable effect, take advantage in real time process.For the realization that legacy detects, existing diverse ways at present.Method one: based on the method for two background modeling, the method sets up the background model of two different update speed, obtains two class prospects by Background difference, comprises the target and at a slow speed or temporarily static target of rapid movement, and then filters out and leave over target; The shortcoming of the method be double-background model to set up computation complexity higher, detection speed is slow, is unsuitable for carrying out in real time leaving over detection.Method two: carry out background modeling and determine target area, and generate suspicious object according to the foreground pixel life period of target area, bianry image according to suspicious object extracts object mask, again specificity analysis is carried out to the image of current scene, utilize the characteristic degree of correlation of the mask image of current scene image and extraction to determine whether as legacy; The method Problems existing is that the last target of reporting to the police of differentiation is not that object is left over or the original allochthonous situation of object in scene further.Method three: the foreground target obtained based on the background modeling method of local updating and the Three image difference of improvement is compared, split the temporary transient resting agglomerate obtained in scene, adopt barycenter to sentence the quiescent time adding up each agglomerate apart from method, judge further to filter out legacy to the static agglomerate reaching time threshold; The shortcoming of the method is that be subject to the impact that pedestrian is blocked, make the time threshold not reaching setting quiescent time added up, after certain hour, legacy incorporates background model, is just no longer detected thus causes undetected when statistics agglomerate quiescent time.
(3) summary of the invention
The present invention is in order to make up the deficiencies in the prior art, provide a kind of remnant object detection method, avoid the problems such as the detection speed using double-background model to cause is slow, complexity is high, target occlusion to a certain degree can be tackled thus reduce the undetected of legacy, decrease the flase drop because the problem such as shade, illumination causes, the object eliminated in scene is moved the flase drop caused, and high problem is failed to report in the wrong report solving legacy detection existence, improves the accuracy that legacy detects.
The present invention is achieved through the following technical solutions:
A kind of remnant object detection method, is characterized in that: comprise the following steps:
(1) image, according to video camera taken, obtains the image of front-end collection;
(2), the image obtained is carried out to the extraction of foreground target;
(3), according to foreground target feature extraction suspicious leave over target;
(4), to the suspicious object extracted carry out signature analysis, tentatively screen out non-legacy target;
(5), to target after screening carry out further signature analysis, determine to leave over target.
Preferably, in step (2), the image of acquisition is carried out gray processing process, obtain its gray level image; Sport foreground is extracted to the gray level image obtained.
Preferably, in step (2), connected domain detection is carried out to the sport foreground extracted; Extract the boundary rectangle of each connected domain, it can be used as moving target; Preserve the barycenter of each moving target, and corresponding timer is set for each moving target.
Preferably, in step (3), target static is for a long time extracted, saved as and suspicious left over target, preserve the image information that this target corresponds to same area in present frame gray figure and background simultaneously, for the suspicious object Offered target degree of correlation with leave over confidence level, and carry out initialization.
Preferably, in step (4), extract the gray level image of preservation and the textural characteristics of present frame same area gray level image respectively, calculate the characteristic similarity of textural characteristics, the similarity of acquisition and the similarity threshold of setting are compared, according to the comparative result more fresh target degree of correlation, the value according to the target degree of correlation determines whether retain this target.
Preferably, in step (4), extract the background image of preservation and the Gradient Features of present frame same area gray level image respectively, calculate its gradient disparities, the gradient disparities of acquisition and the Grads threshold of setting are compared, upgrade according to comparative result and leave over confidence level, then export and leave over confidence level.
Preferably, in step (5), according to object filtering result, what draw target leaves over confidence level, judges final legacy target by the confidence threshold value of setting.
The invention has the beneficial effects as follows: utilize background subtraction method, set up a background model and extract sport foreground, avoid the problems such as the detection speed using double-background model to cause is slow, complexity is high; After sport foreground being detected, utilize barycenter to sentence and carry out the suspicious extraction leaving over target apart from method, and corresponding timer is set for each suspicious target of leaving over, the foreground object detected is represented with barycenter, compare with the barycenter of present frame barycenter and preservation, and the value of timer is adjusted according to comparative result, this process can be tackled target occlusion to a certain degree thus reduce the undetected of legacy; By detect suspicious leave over region and leave over time scene in this region carry out similarity and judge to carry out object filtering further, decrease the flase drop because the problem such as shade, illumination causes; After object filtering, extract the edge feature in this region after leaving over, carry out contrast with the same area edge feature of scene when leaving over and judge, the object eliminated in scene is moved the flase drop caused, thus improves the accuracy of legacy detection.
(4) accompanying drawing explanation
Below in conjunction with accompanying drawing, the present invention is further illustrated.
Accompanying drawing 1 is workflow diagram of the present invention;
Accompanying drawing 2 is the workflow diagram of extraction foreground target of the present invention;
Accompanying drawing 3 is the suspicious workflow diagram leaving over target of extraction of the present invention;
Accompanying drawing 4 is the workflow diagram screening out non-legacy target of the present invention;
Accompanying drawing 5 is the workflow diagram determining to leave over target of the present invention;
(5) embodiment
Accompanying drawing is a kind of specific embodiment of the present invention.This embodiment comprises the following steps: (1), the image taken according to video camera, obtain the image of front-end collection; (2), the image obtained is carried out to the extraction of foreground target; (3), according to foreground target feature extraction suspicious leave over target; (4), to the suspicious object extracted carry out signature analysis, tentatively screen out non-legacy target; (5), to target after screening carry out further signature analysis, determine to leave over target.In step (2), the image of acquisition is carried out gray processing process, obtain its gray level image; Sport foreground is extracted to the gray level image obtained.In step (2), connected domain detection is carried out to the sport foreground extracted; Extract the boundary rectangle of each connected domain, it can be used as moving target; Preserve the barycenter of each moving target, and corresponding timer is set for each moving target.In step (3), extract target static for a long time, saved as and suspicious leave over target, preserve the image information that this target corresponds to same area in present frame gray figure and background simultaneously, for the suspicious object Offered target degree of correlation and leave over confidence level, and carry out initialization.In step (4), extract the gray level image of preservation and the textural characteristics of present frame same area gray level image respectively, calculate the characteristic similarity of textural characteristics, the similarity of acquisition and the similarity threshold of setting are compared, according to the comparative result more fresh target degree of correlation, the value according to the target degree of correlation determines whether retain this target.In step (4), extract the background image of preservation and the Gradient Features of present frame same area gray level image respectively, calculate its gradient disparities, the gradient disparities of acquisition and the Grads threshold of setting are compared, upgrade according to comparative result and leave over confidence level, then export and leave over confidence level.In step (5), according to object filtering result, what draw target leaves over confidence level, judges final legacy target by the confidence threshold value of setting.
Adopt a kind of remnant object detection method of the present invention, its concrete steps are as follows:
Step 101, according to the image of video camera shooting, obtains the image of front-end collection.
Step 102, carries out the extraction of foreground target to the video image obtained.
Step 103, according to foreground target feature extraction suspicious leave over target.
Step 104, carries out signature analysis to the suspicious object extracted, tentatively screens out non-legacy target.
Step 105, to the further signature analysis of target after screening, thus determines to leave over target.
In step 102, the video image of acquisition is carried out gray processing process, obtain its gray level image; Background subtraction method is adopted to detect sport foreground to the gray level image obtained; Connected domain detection is carried out to the sport foreground extracted; Extract the boundary rectangle of each connected domain, it can be used as moving target; Preserve the barycenter of each moving target, and corresponding timer is set for each moving target.
Step 201, carries out gray processing process to the video image of front-end collection, obtains its gray level image.If the pending video of the yuv format obtained, only takes out its Y-component.
Step 202, the gray level image obtained above-mentioned process adopts background subtraction method to detect sport foreground, and namely by setting up background model, analyze each frame with this model bias to detect moving target, generation comprises the bianry image of moving target.
Step 203, application connected domain extraction algorithm, carries out connected component labeling to prospect.
Step 204, adopts the algorithm obtaining image-region minimum enclosed rectangle, extracts the boundary rectangle of each connected domain, gained boundary rectangle is carried out subsequent treatment as moving target.
Step 205, preserves the barycenter of each moving target, i.e. the barycenter of the boundary rectangle of the connected domain of above-mentioned acquisition, and arranges corresponding timer for each moving target.
In step 103, the moving target obtain present frame and the moving target of preservation carry out following process:
Judge the match condition of the moving target of the moving target that present frame obtains and preservation.If coupling, then target timer increases a fixed value.If the moving target that present frame obtains does not mate with the moving target of preservation, then this target is saved as new moving target.
The moving target preserved is processed.If target timer is less than 0, then delete this target of preservation; If target timer is greater than the time threshold of setting, then this target is saved as and suspicious leave over target, preserve the image information that this target corresponds to same area in present frame gray figure and background simultaneously, and for the suspicious object Offered target degree of correlation with leave over confidence level, and carry out initialization.
Step 301, the barycenter of each moving target preserved in obtaining step 102 and timer information thereof.
Step 302, judges the match condition of the moving target of the moving target that present frame obtains and preservation.Employing barycenter sentences the distance between the barycenter of the barycenter of distance method calculating current frame motion target and the moving target of preservation.
Step 303, if the distance between the barycenter of above-mentioned calculating is less than the distance threshold of setting, is then judged to be object matching; If the distance between barycenter is not less than the distance threshold of this setting, be then judged to not mate.
Step 304, above-mentioned be judged to be object matching time, corresponding target timer adds a fixed value a.
Step 305, judges target timer.
Step 306, if the value of target timer is less than 0, then deletes this target.
Step 307, if the value of target timer is not less than 0, judges whether target timer is greater than the time threshold of setting further.
Step 308, if target timer is greater than the time threshold of setting, is then saved as and is suspiciously left over target, and preserves the image information that this target corresponds to same area in present frame gray figure and background.
Step 309, if target timer is greater than the time threshold of setting, preserved suspicious leave over target time, the Offered target degree of correlation and leave over confidence level, and be its initialization.
Step 310, if target timer is not more than the time threshold of setting, is deducted a fixed value b.
In step 104, following process is carried out to the suspicious target of leaving over of preserving:
Extract the gray level image of preservation and the textural characteristics of present frame same area gray level image respectively, calculate the characteristic similarity of textural characteristics, the similarity of acquisition and the similarity threshold of setting are compared, according to the comparative result more fresh target degree of correlation, the value according to the target degree of correlation determines whether retain this target.
Simultaneously, extract the background image of preservation and the Gradient Features of present frame same area gray level image respectively, calculate its gradient disparities, the gradient disparities of acquisition and the Grads threshold of setting are compared, upgrade according to comparative result and leave over confidence level, final output is left over confidence level and is done subsequent treatment.
Step 401, in obtaining step 103, the suspicious of preservation is left over target and corresponding image information, the target degree of correlation and leaves over confidence level.
Step 402, judges whether correspond to suspicious object position in this frame has sport foreground.If there is sport foreground, then do not process; Otherwise carry out following steps.
Step 403, extracts the gray level image of preservation and the textural characteristics of present frame same area gray level image respectively.For LBP textural characteristics: the LBP textural characteristics calculating gray level image and the present frame same area gray level image preserved, statistics LBP feature histogram.
Step 404, carries out Similarity Measure by the textural characteristics of above-mentioned acquisition, obtains the similarity of image.Calculate two histogrammic similarities, computing formula is as follows:
Wherein, i represents Gray Histogram value.
Step 405, compares the similarity of above-mentioned calculating and the similarity threshold of setting.
Step 406, if calculating gained similarity is greater than the similarity threshold of setting, then adds a fixed value c by the target degree of correlation.
Step 407, if calculating gained similarity is not more than the similarity threshold of setting, then subtracts a fixed value d by the target degree of correlation.
Step 408, compares the relevance threshold of the above-mentioned gained target degree of correlation and setting.
Step 409, if the gained target degree of correlation is not more than the relevance threshold of setting, then deletes this suspicious target of leaving over.
Step 410, if the gained target degree of correlation is greater than the relevance threshold of setting, then retains this target and does subsequent treatment.
Step 411, extracts the background image of preservation and the Gradient Features of present frame same area gray level image respectively.Be characterized as example with Sobel, extract the Sobel feature of background image and the current frame image same area gray level image preserved, and by its binary conversion treatment.
Step 412, calculates the gradient disparities preserving background image and present frame gray image.Non-zero pixels number in characteristic pattern after statistics binaryzation, asks the difference of its non-zero pixels number.
Step 413, compares the Grads threshold of above-mentioned gained gradient disparities and setting.
Step 414, if gained gradient disparities is greater than the Grads threshold of setting, then adds 1 by the confidence level of leaving over of this target.
Step 415, if gained gradient disparities is not more than the Grads threshold of setting, then subtracts 1 by the confidence level of leaving over of this target.
Step 416, the confidence level of leaving over exporting target does subsequent treatment.
In step 105, according to the object filtering result of step 104 gained, and target leave over confidence level, by setting confidence threshold value judge final legacy target.
Step 501, the result after obtaining step 104 object filtering.
Step 502, the objective degrees of confidence that obtaining step 104 exports.
Step 503, compares the objective degrees of confidence of acquisition and the confidence threshold value of setting.
Step 504, if objective degrees of confidence is not more than the confidence threshold value of setting, then deletes this target.
Step 505, if objective degrees of confidence is greater than the confidence threshold value of setting, then by this target discrimination for leaving over target.
A series of detailed description listed by the present invention; only illustrating feasibility execution mode of the present invention; and be not used to limit the scope of the invention; when not deviating from the present invention's spirit or essential characteristic; in other specific forms, equivalent way, alter mode realize the present invention, all should be included within protection scope of the present invention.
The present invention describes according to the mode of embodiment, but be not that each execution mode only comprises an independently technical scheme, also should by specification integrally, the technical scheme in each embodiment also through appropriately combined, can form other execution modes that it will be appreciated by those skilled in the art that.
In addition, embodiment of the present invention flow chart and/or block diagram describe, computer program instructions realization flow figure and/or block diagram, except can except supplying method, system (device) or computer program, computer program instructions also can be provided in computer Embedded Processor or other programmable data processing device, make it produce function in flow chart and/or block diagram.
Claims (7)
1. a remnant object detection method, is characterized in that: comprise the following steps:
(1) image, according to video camera taken, obtains the image of front-end collection;
(2), the image obtained is carried out to the extraction of foreground target;
(3), according to foreground target feature extraction suspicious leave over target;
(4), to the suspicious object extracted carry out signature analysis, screen out non-legacy target;
(5), to target after screening carry out further signature analysis, determine to leave over target.
2. a kind of remnant object detection method according to claim 1, is characterized in that: in step (2), the image of acquisition is carried out gray processing process, obtain its gray level image; Sport foreground is extracted to the gray level image obtained.
3. a kind of remnant object detection method according to claim 1, is characterized in that: in step (2), carries out connected domain detection to the sport foreground extracted; Extract the boundary rectangle of each connected domain, it can be used as moving target; Preserve the barycenter of each moving target, and corresponding timer is set for each moving target.
4. a kind of remnant object detection method according to claim 1, it is characterized in that: in step (3), extract target static for a long time, saved as and suspicious left over target, preserve the image information that this target corresponds to same area in present frame gray figure and background simultaneously, for the suspicious object Offered target degree of correlation and leave over confidence level, and carry out initialization.
5. a kind of remnant object detection method according to claim 1, it is characterized in that: in step (4), extract the gray level image of preservation and the textural characteristics of present frame same area gray level image respectively, calculate the characteristic similarity of textural characteristics, the similarity of acquisition and the similarity threshold of setting are compared, according to the comparative result more fresh target degree of correlation, the value according to the target degree of correlation determines whether retain this target.
6. a kind of remnant object detection method according to claim 1, it is characterized in that: in step (4), extract the background image of preservation and the Gradient Features of present frame same area gray level image respectively, calculate its gradient disparities, the gradient disparities of acquisition and the Grads threshold of setting are compared, upgrade according to comparative result and leave over confidence level, then export and leave over confidence level.
7. a kind of remnant object detection method according to claim 1, is characterized in that: in step (5), according to object filtering result, what draw target leaves over confidence level, judges final legacy target by the confidence threshold value of setting.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510985628.6A CN105554462B (en) | 2015-12-25 | 2015-12-25 | A kind of remnant object detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510985628.6A CN105554462B (en) | 2015-12-25 | 2015-12-25 | A kind of remnant object detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105554462A true CN105554462A (en) | 2016-05-04 |
| CN105554462B CN105554462B (en) | 2018-09-21 |
Family
ID=55833373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510985628.6A Active CN105554462B (en) | 2015-12-25 | 2015-12-25 | A kind of remnant object detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105554462B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108572734A (en) * | 2018-04-23 | 2018-09-25 | 哈尔滨拓博科技有限公司 | A kind of gestural control system based on infrared laser associated image |
| CN110163029A (en) * | 2018-02-11 | 2019-08-23 | 中兴飞流信息科技有限公司 | A kind of image-recognizing method, electronic equipment and computer readable storage medium |
| CN110321808A (en) * | 2019-06-13 | 2019-10-11 | 浙江大华技术股份有限公司 | Residue and robber move object detecting method, equipment and storage medium |
| CN110490025A (en) * | 2018-05-14 | 2019-11-22 | 杭州海康威视数字技术股份有限公司 | A kind of object detection method, device, equipment and system |
| CN110751107A (en) * | 2019-10-23 | 2020-02-04 | 北京精英系统科技有限公司 | Method for detecting event of discarding articles by personnel |
| CN110991370A (en) * | 2019-12-09 | 2020-04-10 | 浩云科技股份有限公司 | Multi-channel information fusion ATM panel remnant detection method |
| CN113470013A (en) * | 2021-07-28 | 2021-10-01 | 浙江大华技术股份有限公司 | Method and device for detecting moved article |
| CN113743212A (en) * | 2021-08-02 | 2021-12-03 | 日立楼宇技术（广州）有限公司 | Detection method and device for jam or left object at entrance and exit of escalator and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299274A (en) * | 2008-06-18 | 2008-11-05 | 北京中星微电子有限公司 | Detecting method and system for moving fixed target |
| CN101727672A (en) * | 2008-10-24 | 2010-06-09 | 云南正卓信息技术有限公司 | Method for detecting, tracking and identifying object abandoning/stealing event |
| US20120062738A1 (en) * | 2009-05-19 | 2012-03-15 | Panasonic Corporation | Removal/abandonment determination device and removal/abandonment determination method |
| CN103729858A (en) * | 2013-12-13 | 2014-04-16 | 广州中国科学院先进技术研究所 | Method for detecting article left over in video monitoring system |
| CN104156942A (en) * | 2014-07-02 | 2014-11-19 | 华南理工大学 | Detection method for remnants in complex environment |
-
2015
- 2015-12-25 CN CN201510985628.6A patent/CN105554462B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299274A (en) * | 2008-06-18 | 2008-11-05 | 北京中星微电子有限公司 | Detecting method and system for moving fixed target |
| CN101727672A (en) * | 2008-10-24 | 2010-06-09 | 云南正卓信息技术有限公司 | Method for detecting, tracking and identifying object abandoning/stealing event |
| US20120062738A1 (en) * | 2009-05-19 | 2012-03-15 | Panasonic Corporation | Removal/abandonment determination device and removal/abandonment determination method |
| CN103729858A (en) * | 2013-12-13 | 2014-04-16 | 广州中国科学院先进技术研究所 | Method for detecting article left over in video monitoring system |
| CN104156942A (en) * | 2014-07-02 | 2014-11-19 | 华南理工大学 | Detection method for remnants in complex environment |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163029A (en) * | 2018-02-11 | 2019-08-23 | 中兴飞流信息科技有限公司 | A kind of image-recognizing method, electronic equipment and computer readable storage medium |
| CN110163029B (en) * | 2018-02-11 | 2021-03-30 | 中兴飞流信息科技有限公司 | Image recognition method, electronic equipment and computer readable storage medium |
| CN108572734A (en) * | 2018-04-23 | 2018-09-25 | 哈尔滨拓博科技有限公司 | A kind of gestural control system based on infrared laser associated image |
| CN110490025A (en) * | 2018-05-14 | 2019-11-22 | 杭州海康威视数字技术股份有限公司 | A kind of object detection method, device, equipment and system |
| CN110490025B (en) * | 2018-05-14 | 2023-05-05 | 杭州海康威视数字技术股份有限公司 | Target detection method, device, equipment and system |
| CN110321808A (en) * | 2019-06-13 | 2019-10-11 | 浙江大华技术股份有限公司 | Residue and robber move object detecting method, equipment and storage medium |
| CN110321808B (en) * | 2019-06-13 | 2021-09-14 | 浙江大华技术股份有限公司 | Method, apparatus and storage medium for detecting carry-over and stolen object |
| CN110751107A (en) * | 2019-10-23 | 2020-02-04 | 北京精英系统科技有限公司 | Method for detecting event of discarding articles by personnel |
| CN110991370A (en) * | 2019-12-09 | 2020-04-10 | 浩云科技股份有限公司 | Multi-channel information fusion ATM panel remnant detection method |
| CN110991370B (en) * | 2019-12-09 | 2023-07-28 | 浩云科技股份有限公司 | Multichannel information fusion ATM panel carryover detection method |
| CN113470013A (en) * | 2021-07-28 | 2021-10-01 | 浙江大华技术股份有限公司 | Method and device for detecting moved article |
| CN113743212A (en) * | 2021-08-02 | 2021-12-03 | 日立楼宇技术（广州）有限公司 | Detection method and device for jam or left object at entrance and exit of escalator and storage medium |
| CN113743212B (en) * | 2021-08-02 | 2023-11-14 | 日立楼宇技术（广州）有限公司 | Method and device for detecting congestion or carryover at entrance and exit of escalator and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105554462B (en) | 2018-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105554462A (en) | Remnant detection method | |
| WO2021012757A1 (en) | Real-time target detection and tracking method based on panoramic multichannel 4k video images | |
| Li et al. | Real-time moving vehicle detection, tracking, and counting system implemented with OpenCV | |
| CN104933710B (en) | Based on the shop stream of people track intelligent analysis method under monitor video | |
| CN103246896B (en) | A kind of real-time detection and tracking method of robustness vehicle | |
| CN102222214A (en) | Fast object recognition algorithm | |
| Karpagavalli et al. | Estimating the density of the people and counting the number of people in a crowd environment for human safety | |
| CN106296677A (en) | A kind of remnant object detection method of double mask context updates based on double-background model | |
| CN103729858A (en) | Method for detecting article left over in video monitoring system | |
| CN104616006A (en) | Surveillance video oriented bearded face detection method | |
| CN103886609B (en) | Vehicle tracking method based on particle filtering and LBP features | |
| Aoun et al. | Graph modeling based video event detection | |
| CN103577832A (en) | People flow statistical method based on spatio-temporal context | |
| Almomani et al. | Segtrack: A novel tracking system with improved object segmentation | |
| Jin et al. | Fusing Canny operator with vibe algorithm for target detection | |
| CN106295523A (en) | A kind of public arena based on SVM Pedestrian flow detection method | |
| CN116453069A (en) | Expressway casting object detection method based on cascade difference perception model | |
| Kaur et al. | An Efficient Method of Number Plate Extraction from Indian Vehicles Image | |
| Yu et al. | Length-based vehicle classification in multi-lane traffic flow | |
| CN115188081A (en) | Complex scene-oriented detection and tracking integrated method | |
| Qin et al. | A background extraction and shadow removal algorithm based on clustering for ViBe | |
| Zhen-Jie et al. | Research on Detection and Tracking of Moving Vehicles in Complex Environment Based on Real-Time Surveillance Video | |
| CN106874864A (en) | A kind of outdoor pedestrian's real-time detection method | |
| CN107590418A (en) | A kind of video smoke recognition methods based on behavioral characteristics | |
| Suresh et al. | A survey on occlusion 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200107 Address after: 250000 East 5-301, Yinhe building, No. 2008, Xinluo street, high tech Zone, Jinan City, Shandong Province Patentee after: Jinan Aiwei Internet Co., Ltd Address before: Xinluo Avenue high tech Zone of Ji'nan City, Shandong province 250000 silver bearing No. 2008 Building No. 4-1101 Patentee before: Jovision Technology Co., Ltd. |
|
| TR01 | Transfer of patent right |