WO1998028706B1 - Low false alarm rate video security system using object classification - Google Patents
Low false alarm rate video security system using object classificationInfo
- Publication number
- WO1998028706B1 WO1998028706B1 PCT/US1997/024163 US9724163W WO9828706B1 WO 1998028706 B1 WO1998028706 B1 WO 1998028706B1 US 9724163 W US9724163 W US 9724163W WO 9828706 B1 WO9828706 B1 WO 9828706B1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- scene
- segment
- image
- segments
- security system
- Prior art date
Links
- 230000000694 effects Effects 0.000 claims abstract 6
- 238000000034 method Methods 0.000 claims abstract 3
- 230000000875 corresponding Effects 0.000 claims 7
- 238000000605 extraction Methods 0.000 claims 5
- 238000003384 imaging method Methods 0.000 claims 1
- 230000000873 masking Effects 0.000 claims 1
- 241001465754 Metazoa Species 0.000 abstract 2
- 238000001514 detection method Methods 0.000 abstract 1
Abstract
A video detection system (10) and method detects an intruder from video images of a scene. The method employs a recognition process to differentiate between humans and animals. The method is used only after possible false alarms resulting from identified effects of noise, aliasing, non-intruder motion occuring within the scene, and effects of global or local lighting changes. The object recognition process includes determining the regions containing a potential intruder, outlining and growing those regions to encompass all of potential intruders, determining a set of shape features from the region and eliminating possible shadow effects, normalizing the set, and comparing the normalized set with sets of features of humans and animals. This comparison produces a confidence level indicating a human intruder. An alarm is given for a sufficiently high confidence level. The possibility of a false alarm due to an animal or a non-identifiable object is also substantially eliminated.
Claims
1. A video security system visually monitoring a scene and detecting motion of an object within the scene comprising: imaging means continually viewing the scene and producing a signal representative of the scene; processor means processing said signal, comparing the signal representing the scene at one point in time with a similar signal representing the scene at a previous point in time, and identifying those segments of the scene at said point in time which differs from segments of the scene at the earlier points in time; discriminator means evaluating those segments of the scene identified as being different to determine if the differences are caused by surface differences which are indicative of the presence of an intruder within the scene, or lighting changes which occur within the scene and do not indicate the presence of an intruder, if the difference is caused by the presence of an intruder providing an indication thereof; said discriminator means including means comparing pixel elements contained in each segment of the scene at one point in time and corresponding pixel elements contained in a corresponding segment from the scene at an earlier point in time; said discriminator means determining a ratio of light intensity between each pixel in a segment with each pixel adjacent thereto; and means comparing the ratio values for the pixels in the segment of the scene at one point in time with the ratio values for the pixels in the corresponding segment of the scene at the earlier point in time; and, said discriminator means further evaluating those segments of the scene identified as being different to classify each segment as a human life form or not, to give an alarm whenever an object present in one of the segments is classified as a human life form representing a human intruder within the scene, and to give no alarm if objects present in the segments are classified as non-human life forms.
2. The video security system of claim 1 wherein said discriminator means further includes threshold means determining if a computed ratio for one pixel and an adjacent pixel in the segment of the scene at the one point in time varies by a predetermined threshold value from the computed ratio for the corresponding pixels in the scene segment at the earlier point in time.
3. The video security system of claim 2 wherein said discriminator means further includes means growing each segment to a size which incorporates all of the pixels which define an object contained within the segment.
4. The video security system of claim 3 wherein said discriminator means further includes means extracting a set of features from the object.
5. The video security system of claim 4 wherein said feature extraction means includes means extracting linear shape features from the object as numerical values representing such factors as the height, width, horizontal and vertical edges of the object, and degree of circularity of the object.
6. The video security system of claim 4 wherein said feature extraction means further includes means extracting Fourier descriptors of the silhouette shape features of the object.
7. The video security system of claim 6 wherein said feature extraction means further includes means normalizing any value obtained from the feature extraction means in the event said value falls outside a predetermined range of values for the particular feature.
8. The video security system of claim 6 wherein said discriminator means further includes classifier means evaluating said set of features for said object with sets of features representing human and non-human life forms and for deriving a value representing a degree of confidence as to the correspondence of the object to a human or non-human life form.
9. The video security system of claim 8 further including means providing an alarm indication only if the degree of confidence for the correspondence of the object to a human life form exceeds a predetermined confidence level.
10. The video security system of claim 8 wherein said classifier means includes a linear object classification means providing a confidence level output for each of the three classes: human, animal, and unknown. - 29 -
11. The video security system of claim 8 wherein said classifier means includes a non-linear object classification means providing a confidence level output for each of three classes: human, animal, and unknown.
12. The video security system of claim 1 wherein said discrimination means includes means executing an algorithm to perform object classification.
13. The video security system of claim 4 wherein said feature extraction means includes means eliminating shadows cast by an object represented by the segment.
14. The video security system of claim 9 wherein said alarm indication means further provides a second alarm indication if an object is classified as unknown.
15. A method of evaluating a scene to determine if any perceived movement within the scene is caused by an intruder into the scene comprising: viewing the scene and creating an image of the scene comprising a plurality of pixels arranged in an array; comparing the image of the scene with a reference image thereof to produce a difference image, producing said difference image including convolving the image with an antialiasing means to eliminate any aliasing effects in the difference image, outlining any segments where possible movement has occurred, determining a ratio of light intensity between each pixel in a segment with each pixel adjacent thereto, and comparing the ratio values for the pixels in a segment of one image with the ratio values for the pixels in the corresponding segment of another image to eliminate the effects of lighting changes; processing the difference image to identify any segments therewithin which, based upon a first predetermined set of criteria, represent spatially constrained movements of an object fixed within the scene, and further processing the difference image to identify any segments therewithin which, based upon a second predetermined set of criteria, represent artifacts not caused by the presence of an intruder within the scene, said segments meeting said first and second sets of criteria being identified as segments not requiring further - 30 -
processing; and, further processing those segments within the difference image which remain to determine if movement therewithin is caused by an intruder.
16. The method of claim 15 wherein viewing the scene includes continually viewing the scene with a camera producing a two-dimensional image of the scene.
17. The method of claim 15 wherein processing said difference image includes region growing the segments within the difference image.
18. The method of claim 17 wherein processing said difference image in accordance with said second predetermined set of criteria further includes determining if each region grown segment within the image in which an artifact occurs is less than a predetermined size and identifying those segments whose size is less than the predetermined size as resulting from such causes as noise or lighting effects, and not as the result of an intruder.
19. The method of claim 15 wherein processing the difference image and identifying segments therewithin representing spatially constrained movements of an object fixed within the scene further includes masking those segments within the scene in which said movements are confined.
20. The method of claim 19 further including processing those masked segments in the scene when a segment in which an intruder may be present overlaps a boundary of a masked segment, but to not otherwise process said masked segments.
21. The method claim 15 further including determining if a computed ratio for one pixel and an adjacent pixel in one segment of an image differs by a predetermined threshold value from the computed ratio for the pixels in the corresponding segment of the other image.
22. The method of claim 21 further including determining if the number of computed ratios for the pixels in the segment of the one image compared with the computed ratio for the pixels in the corresponding segment of the other image exceed a second threshold value, the result of the comparison, if - 31 -
exceeding the second threshold value, indicating that the difference between the images is the result of the presence of an intruder introduced into the scene.
23. The method of claim 15 further including updating the reference image of the scene if the comparison of an image with a reference image reveals no possible motion within the scene has occurred.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CA002275893A CA2275893C (en) | 1996-12-23 | 1997-12-23 | Low false alarm rate video security system using object classification |
EP97954298A EP1010130A4 (en) | 1996-12-23 | 1997-12-23 | Low false alarm rate video security system using object classification |
AU58109/98A AU5810998A (en) | 1996-12-23 | 1997-12-23 | Low false alarm rate video security system using object classification |
Applications Claiming Priority (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US77199196A | 1996-12-23 | 1996-12-23 | |
US08/772,595 | 1996-12-23 | ||
US08/772,595 US5937092A (en) | 1996-12-23 | 1996-12-23 | Rejection of light intrusion false alarms in a video security system |
US08/771,991 | 1996-12-23 | ||
US08/772,731 US5956424A (en) | 1996-12-23 | 1996-12-23 | Low false alarm rate detection for a video image processing based security alarm system |
US08/772,731 | 1996-12-23 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO1998028706A1 WO1998028706A1 (en) | 1998-07-02 |
WO1998028706B1 true WO1998028706B1 (en) | 1998-09-11 |
Family
ID=27419676
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US1997/024163 WO1998028706A1 (en) | 1996-12-23 | 1997-12-23 | Low false alarm rate video security system using object classification |
Country Status (4)
Country | Link |
---|---|
EP (1) | EP1010130A4 (en) |
AU (1) | AU5810998A (en) |
CA (1) | CA2275893C (en) |
WO (1) | WO1998028706A1 (en) |
Families Citing this family (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
NO982640L (en) * | 1998-06-08 | 1999-12-09 | Nyfotek As | Method and system for monitoring an area |
GB9822956D0 (en) | 1998-10-20 | 1998-12-16 | Vsd Limited | Smoke detection |
ATE250260T1 (en) * | 1999-07-17 | 2003-10-15 | Siemens Building Tech Ag | ROOM MONITORING DEVICE |
US6774905B2 (en) | 1999-12-23 | 2004-08-10 | Wespot Ab | Image data processing |
US6819353B2 (en) | 1999-12-23 | 2004-11-16 | Wespot Ab | Multiple backgrounds |
US7479980B2 (en) | 1999-12-23 | 2009-01-20 | Wespot Technologies Ab | Monitoring system |
SE519700C2 (en) * | 1999-12-23 | 2003-04-01 | Wespot Ab | Image Data Processing |
SE517900C2 (en) * | 1999-12-23 | 2002-07-30 | Wespot Ab | Methods, monitoring system and monitoring unit for monitoring a monitoring site |
US6940998B2 (en) | 2000-02-04 | 2005-09-06 | Cernium, Inc. | System for automated screening of security cameras |
GB0028162D0 (en) * | 2000-11-20 | 2001-01-03 | Sentec Ltd | Distributed image processing technology and services |
US7212651B2 (en) * | 2003-06-17 | 2007-05-01 | Mitsubishi Electric Research Laboratories, Inc. | Detecting pedestrians using patterns of motion and appearance in videos |
EP1672604A1 (en) * | 2004-12-16 | 2006-06-21 | Siemens Schweiz AG | Method and apparatus for detection of tampering with a surveillance camera |
US7822224B2 (en) | 2005-06-22 | 2010-10-26 | Cernium Corporation | Terrain map summary elements |
US7526105B2 (en) * | 2006-03-29 | 2009-04-28 | Mark Dronge | Security alarm system |
ATE521054T1 (en) | 2006-12-20 | 2011-09-15 | Axis Ab | METHOD AND DEVICE FOR DETECTING SABOTAGE ON A SURVEILLANCE CAMERA |
WO2010124062A1 (en) | 2009-04-22 | 2010-10-28 | Cernium Corporation | System and method for motion detection in a surveillance video |
CN102169614B (en) * | 2011-01-14 | 2013-02-13 | 云南电力试验研究院(集团)有限公司 | Monitoring method for electric power working safety based on image recognition |
CN106878668B (en) | 2015-12-10 | 2020-07-17 | 微软技术许可有限责任公司 | Movement detection of an object |
US10535252B2 (en) | 2016-08-10 | 2020-01-14 | Comcast Cable Communications, Llc | Monitoring security |
GB2557597B (en) * | 2016-12-09 | 2020-08-26 | Canon Kk | A surveillance apparatus and a surveillance method for indicating the detection of motion |
CN110113561A (en) * | 2018-02-01 | 2019-08-09 | 广州弘度信息科技有限公司 | A kind of personnel are detained detection method, device, server and system |
JP7415872B2 (en) * | 2020-10-23 | 2024-01-17 | 横河電機株式会社 | Apparatus, system, method and program |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2503613B2 (en) * | 1988-12-23 | 1996-06-05 | 松下電工株式会社 | Abnormality monitoring device |
US5144685A (en) * | 1989-03-31 | 1992-09-01 | Honeywell Inc. | Landmark recognition for autonomous mobile robots |
US5274714A (en) * | 1990-06-04 | 1993-12-28 | Neuristics, Inc. | Method and apparatus for determining and organizing feature vectors for neural network recognition |
US5493273A (en) * | 1993-09-28 | 1996-02-20 | The United States Of America As Represented By The Secretary Of The Navy | System for detecting perturbations in an environment using temporal sensor data |
JPH07192112A (en) * | 1993-12-27 | 1995-07-28 | Oki Electric Ind Co Ltd | Intruding object recognizing method |
-
1997
- 1997-12-23 CA CA002275893A patent/CA2275893C/en not_active Expired - Lifetime
- 1997-12-23 WO PCT/US1997/024163 patent/WO1998028706A1/en not_active Application Discontinuation
- 1997-12-23 AU AU58109/98A patent/AU5810998A/en not_active Abandoned
- 1997-12-23 EP EP97954298A patent/EP1010130A4/en not_active Withdrawn
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO1998028706B1 (en) | Low false alarm rate video security system using object classification | |
US6104831A (en) | Method for rejection of flickering lights in an imaging system | |
US8189049B2 (en) | Intrusion alarm video-processing device | |
US5956424A (en) | Low false alarm rate detection for a video image processing based security alarm system | |
CN105354563B (en) | Face datection prior-warning device and implementation method are blocked in conjunction with depth and color image | |
JP3123587B2 (en) | Moving object region extraction method using background subtraction | |
US5937092A (en) | Rejection of light intrusion false alarms in a video security system | |
CN111368771A (en) | Tunnel fire early warning method and device based on image processing, computer equipment and computer readable storage medium | |
CA2275893C (en) | Low false alarm rate video security system using object classification | |
US20060170769A1 (en) | Human and object recognition in digital video | |
WO2005122094A1 (en) | Method for detecting desired objects in a highly dynamic environment by a monitoring system | |
JP2010238032A (en) | Smoke detection device | |
KR101030257B1 (en) | Method and System for Vision-Based People Counting in CCTV | |
CN111259718A (en) | Escalator retention detection method and system based on Gaussian mixture model | |
CN108830161B (en) | Smog identification method based on video stream data | |
AU2002232008B2 (en) | Method of detecting a significant change of scene | |
AU2002232008A1 (en) | Method of detecting a significant change of scene | |
JP2612662B2 (en) | Intruder identification system | |
JP2010218046A (en) | Smoke detection device | |
Frejlichowski et al. | Extraction of the foreground regions by means of the adaptive background modelling based on various colour components for a visual surveillance system | |
JP2007060468A (en) | Image sensor | |
JPH05300516A (en) | Animation processor | |
WO2020139071A1 (en) | System and method for detecting aggressive behaviour activity | |
JPH0731248B2 (en) | Moving object detector | |
Freer et al. | Automatic video surveillance with intelligent scene monitoring and intruder detection |