WO2007095160A2 - Smoke detection method and apparatus - Google Patents
Smoke detection method and apparatus Download PDFInfo
- Publication number
- WO2007095160A2 WO2007095160A2 PCT/US2007/003655 US2007003655W WO2007095160A2 WO 2007095160 A2 WO2007095160 A2 WO 2007095160A2 US 2007003655 W US2007003655 W US 2007003655W WO 2007095160 A2 WO2007095160 A2 WO 2007095160A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- identified areas
- smoke
- areas
- bitmaps
- monitored area
- Prior art date
Links
- 239000000779 smoke Substances 0.000 title claims abstract description 88
- 238000001514 detection method Methods 0.000 title description 9
- 230000002123 temporal effect Effects 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims description 20
- 238000004458 analytical method Methods 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 3
- 230000005484 gravity Effects 0.000 claims description 3
- 238000013459 approach Methods 0.000 description 13
- 239000002245 particle Substances 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000000977 initiatory effect Effects 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 239000003570 air Substances 0.000 description 2
- 238000000149 argon plasma sintering Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009792 diffusion process Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 239000000443 aerosol Substances 0.000 description 1
- 239000012080 ambient air Substances 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
- G08B17/125—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions by using a video camera to detect fire or smoke
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Definitions
- the present invention generally relates to electrical, condition responsive systems and methods. More particularly, this invention relates to a method and apparatus for detecting and identifying developing smoke clouds in a monitored area using a sequence of digitized images.
- Smoke detectors are very important safety devices that can provide an early warning of fire in a monitored area. Considerable efforts have been devoted to improving upon the technology used in smoke detectors as a means of increasing their usefulness and reliability.
- One of the and most commonly used methodologies for smoke detectors involves measuring the presence of aerosol particles at the location of a smoke detector's sensor. Such measurements are based either on light scattering phenomena or on the effects due to smoke particle interactions with an ionization current created within the detector. See Rattman , et al., U.S. Patent No. (USPN) 5,719,557.
- a disadvantage of this approach is that its measurements are limited in terms of their sensing area since such detectors monitor for the presence of smoke only at those points that are in close proximity to the location of the detector's sensor.
- the successful detection of smoke in a monitored area using this technique greatly depends upon the rate of movement of smoke particles toward the detector's sensor which, depending upon the size of the monitored area, can be located at considerable distance from the initial source of any smoke.
- air samples be collected at multiple locations in the monitored area and then to guide these samples to the location of the detector's sensor. See Knox, et al., USPN 6,285,291.
- a disadvantage of using such a laser beam approach is that, although it may effectively measure smoke conditions at more points within a monitored area that just those points in the vicinity of a single sensor, it still does not provided feedback on the smoke conditions at all or most of the points within the monitored area.
- multiple practitioners have introduced yet another approach for detecting smoke. Instead of directly measuring the attenuation or scattering of light at the given point in space, their approach uses a digital image that is produced by a TV and identifies the presence of smoke in the viewed area by the effects of the smoke on the video image.
- Grech-Cini' s USPN 6,844,8108 each pixel of the image is constantly analyzed statistically against the average parameters across the image. The characteristics that are monitored include brightness, color intensity and spatial contrast.
- a Bayesian estimator is produced for each pixel and is compared to values from the current image. Once a significant deviation from the estimate occurs over a range of pixels, a determination is made on whether the anomaly is caused by smoke.
- the disadvantage of this approach is 1 that it has been found to yield a high rate of false alarms that may be caused by
- OBJECTS AND ADVANTAGES There has been summarized above, rather broadly, the prior art that is related to the present invention in order that the context of the present invention may be better understood and appreciated. In this regard, it is instructive to also consider the objects and advantages of the present invention. It is an object of the present invention to provide apparatus and methods that are effective at detecting smoke within the entire volume of a monitored area. It is another object of the present invention to provide apparatus and methods that are effective at detecting smoke in industrial petrochemical installations. It is an object of the present invention to provide apparatus and methods that can operate within the framework of the ordinary Closed Circuit Television (CCTV) surveillance systems that are used to monitor commercial, outdoor, industrial and residential areas.
- CCTV Closed Circuit Television
- an apparatus for detecting smoke in a monitored area that includes: (a) a means for capturing a series of successive video images of the monitored area as a series of two-dimensional bitmaps having a specified number of pixels, (b) a means for examining this series of bitmaps to identify areas in successive bitmaps of adjacent pixels that have the potential for being used as indicators for the existence of smoke in the monitored area, (c) a means for storing these identified areas from the bitmaps, and (d) a means for analyzing the identified areas so as to identify the temporal changes in these areas which are indicative of the presence of smoke.
- Alternative embodiments of the present invention can be obtained by adding to this first embodiment any combination of the following elements: (e) a means for analytically characterizing the data contained in each of the identified areas such that the amount of this data is reduced and specifiable information pertaining to these identified areas is preserved so as to allow one to analyze this reduced data to identify the presence of smoke in the monitored area, and (f) a means for generating an alarm when the means for analyzing the identified areas identifies the temporal changes that are indicative of the presence of smoke.
- Still further embodiments of the present invention can be obtained by configuring these means as follows: (g) the means for examining so that it selects those pixels having an approximate uniform change in intensity, (h) the means for analyzing so that it selects those temporal changes chosen from the group consisting of those temporal changes which indicate an ever-increasing number of pixels in those identified areas of the cluster or those temporal changes which indicate that the increase in the size of the identified areas occurs to a greater extent in the top rather than the bottom part of the identified areas, or (i) the means for analyzing so that it makes use of a neural network.
- FIG. 1 shows a block diagram of a preferred embodiment of the smoke detection method and apparatus of the present invention.
- FIG. 2 shows the algorithm of a preferred embodiment of the smoke detection method and apparatus of the present invention.
- FIG. 3 illustrates the particular trends that are typical for a growing smoke plume.
- FIG. 4 illustrates the specific trends that are typical for a walking human.
- FIG. 5 shows a possible conversion of the growing plume into a 2-D pattern.
- FIG. 1 shows a block diagram of a preferred embodiment of the present invention.
- the images of a growing smoke cloud within a specified area are seen to be monitored by a camera 2 that converts the image into a digital representation within the computer memory of a video capturing device 4.
- This representation takes the form of sequence of digital images which are analyzed by a computer system 6.
- the analysis of these images consists of: (a) the identification and extraction of the pixel areas (blobs) that are candidates for being representative of smoke clouds (i.e., these blobs have the potential for being used as indicators of smoke in the monitored area), (b) the tracking of these blobs to identify their long term trends, and (c) comparing these trends to identify those which are consistent with those produced by a growing smoke cloud.
- a prescribed response system 8 is activated.
- the computer system 6 identifies a group of pixels (i.e., blob) that may represent a cloud of smoke by monitoring the changes in the properties of the pixels that comprise the group.
- Such changes may include, but are not limited to, consistent (across the blob) changes in brightness or intensity.
- a cloud of black smoke will make all the pixels within the cloud darker, while white smoke will make all the pixels lighter.
- blobs are arranged in a tracking cluster that stores all the sequence of overlapping blobs from a series of consecutive frames (i.e., a cluster is only the set of overlapping regions or blobs of the bitmaps; not the entire bitmaps themselves).
- FIG. 2 shows a flow-chart representation of such an algorithm.
- the analysis scheme of the present invention begins with the identification of a blob of interest. At its simplest form, this step could be achieved by, for example, searching a captured digital image for a group of pixels with consistent brightness changes.
- the analysis scheme checks as to whether this new image's blob should be chronologically associated, on the basis of the likelihood of the existence of smoke in the monitored area, with any of the blobs that are being tracked from earlier images. This is achieved by checking for spatial overlap between this new blob and the database of tracking clusters identified from the prior images. In case of such overlap, the blob is appended to the appropriate tracking cluster. If no such overlap exists, this blob becomes the first of what may be a sequence of blobs that will make up yet another cluster. Clusters that have no new blobs added to them are assumed to no longer be active and are purged from the database.
- Each updated tracking cluster is analyzed to identify whether the addition of this new blob allows one to identify in such clusters any of the special patterns or signatures of growth or movement that can be associated with a particular type of physical phenomena (e.g., growth and movement of a smoke cloud).
- a prescribed system response can be activated. Meanwhile, this analysis scheme proceeds to analyze the next blob.
- a representative example of the identifiable signatures in the growth and movement of the blobs within the clusters for growing smoke clouds is shown in FIG. 3.
- the contours of the blobs from consecutive video frames 1, 2 and 3 are seen to have spatial overlay.
- the changes in the growth, movement or shape of representative smoke cloud blobs can be caused by several factors, including: (a) constant supply of smoke from a source, (b) diffusion of the smoke, and (c) air movements within the monitored space.
- Examples of the blob movements or size changes that can be used to identify such smoke clouds include: (i) blob boundary expansions as more smoke is generated, (U) blob boundary movement and expansion towards the top of the blob due to the hotter gases that comprise the smoke cloud rising in the surrounding, cooler ambient air, and (iii) some area of blob overlay between consecutive images.
- FIG. 4 shows the evolution of the blob that is caused by a human moving across the field of view of a camera.
- Technical barriers to the implementation of various embodiments of the present invention include the need to accommodate for the "fuzziness" which is usually seen in actual video data.
- One means of addressing this problem is the use of a feed-forward, three layer, neural network (those knowledgeable in such technology will recognize that similar results may be achieved by utilizing different neural network layouts).
- FIG. 5 shows a three dimensional (time plus a two dimensional plane) space for the representation and analysis of such clusters.
- the vertical axis of this space is defined to be the time since a desired sequence of video images began to be accumulated, while its horizontal, X-Y, plane represents the spatial extent of a video image.
- the digital image of a smoke cloud is assumed to be a solid, two dimensional figure (i.e., it has no interior voids; it has a continuous enclosing contour or boundary) such as those represented by one of the slices that are stacked one above another in the temporal plane of FIG. 5.
- the lateral extent of these slices is seen to be increasing in time (i.e., the sizes of the images taken at instants that are a considerable time after the initiation of the smoke cloud are larger than those taken when the smoke cloud had just begun to form).
- FIG. 5 In the lower, left-hand side of FIG. 5 is shown a two-dimensional, normalized (12x12) overlay of these slices (i.e., a footprint) in which each cell of this footprint represents the relative (to age of the cluster) time span that the particular cell has been covered by the smoke cloud, and the phase (time of initiation relative to age of the cluster) when the smoke cloud first covered the cell in question.
- a footprint In the lower, left-hand side of FIG. 5 is shown a two-dimensional, normalized (12x12) overlay of these slices (i.e., a footprint) in which each cell of this footprint represents the relative (to age of the cluster) time span that the particular cell has been covered by the smoke cloud, and the phase (time of initiation relative to age of the cluster) when the smoke cloud first covered the cell in question.
- each cell in this footprint can be represented by a 2-d vector in which the phase, ⁇ , of this vector is defined so as to provide information on the time when the pixel was "energized” or at what exact moment pixel has been assigned smoke status (e.g., ⁇ ranges from 0 to ⁇ /2, with ⁇ /2 being defined as the youngest pixels, or those that have been identified at the greatest time since the initiation of the smoke cloud was recognized).
- the length, ⁇ , of this 2-d vector may be used to encode, for example, the "age of cell,” which is the time duration that it has been covered by the cloud.
- the typical smoke cloud yields a cluster that has its longest vector with close to zero phase at the location of the smoke's source, while shorter and similarly higher phased vectors will be at the peripheral locations of the smoke cloud.
- Various alternative approaches that are considered to come within the scope of the present invention for the identification of a smoke cloud include: (a) trend analysis of a smoke cloud's contour area (e.g., an ever-increasing number of pixels in those identified areas of a cluster, or the identification of cloud's contour area changes as a positive linear trend over a prescribed period of time with the computed least squares value for said trend being below a prescribed level, or the identification of cloud's contour area changes by a polynomial least square approximation with a consistent positive first derivative exceeding a prescribed value over the a prescribed period and with the least square value being below a prescribed level), (b) movement analysis in the top and bottom segments of a cloud contour (e.g., the increase in the size of the identified areas occurs to a greater extent in the top rather than the bottom part of the identified areas, where the top and bottom parts are defined with respect to the direction of influence of gravity), and (c) various other, logical, rule-based, predicative approaches.
- the present invention is implemented on an Intel® PentiumTM IV class processor platform under a Windows operating system.
- This processing system is capable of simultaneously handling eight video channels and can also provide the necessary alarm notifications over EP networks.
- the algorithms of the present invention have been implemented in C++ programming language using the Microsoft Visual Studio .NET environment.
- a feed forward, three-layer neural network (294/20/1) configuration is used for pattern matching. Neural network training for smoke cloud identification was done on over 10,000 samples from which more than 95% are nuisances.
Abstract
An apparatus for detecting smoke in a monitored area includes: (a) a means for capturing a series of successive video images of the monitored area as a series of two-dimensional bitmaps having a specified number of pixels, (b) a means for examining this series of bitmaps to identify areas in successive bitmaps of adjacent pixels that have the potential for being used as indicators for the existence of smoke in the monitored area, (c) a means for storing these identified areas from the bitmaps, (d) a means for associating, on the basis of the likelihood of the existence of smoke in the monitored area, the identified areas from successive bitmaps so as to form a cluster of overlapping regions within those bitmaps, and (d) a means for analyzing the identified areas so as to identify the temporal changes in these areas which are indicative of the presence of smoke.
Description
SMOKE DETECTION METHOD AND APPARATUS
BACKGROUND OF THE INVENTION
FIELD OF THE INVENTION The present invention generally relates to electrical, condition responsive systems and methods. More particularly, this invention relates to a method and apparatus for detecting and identifying developing smoke clouds in a monitored area using a sequence of digitized images.
DESCRIPTION OF THE RELATED ART Smoke detectors are very important safety devices that can provide an early warning of fire in a monitored area. Considerable efforts have been devoted to improving upon the technology used in smoke detectors as a means of increasing their usefulness and reliability. One of the and most commonly used methodologies for smoke detectors involves measuring the presence of aerosol particles at the location of a smoke detector's sensor. Such measurements are based either on light scattering phenomena or on the effects due to smoke particle interactions with an ionization current created within the detector. See Rattman , et al., U.S. Patent No. (USPN) 5,719,557. A disadvantage of this approach is that its measurements are limited in terms of their sensing area since such detectors monitor for the presence of smoke
only at those points that are in close proximity to the location of the detector's sensor. The successful detection of smoke in a monitored area using this technique greatly depends upon the rate of movement of smoke particles toward the detector's sensor which, depending upon the size of the monitored area, can be located at considerable distance from the initial source of any smoke. To address this insufficient sample size problem, it has been suggested that air samples be collected at multiple locations in the monitored area and then to guide these samples to the location of the detector's sensor. See Knox, et al., USPN 6,285,291. Although effectively increasing the extent of spatial sampling within a monitored area, this method has the disadvantage of requiring the installation of multiple sampling tubes at assorted locations throughout the monitored area. Another approach for smoke detection has been to monitor the light scattering effect of smoke particles on a laser beam that is directed across a monitored area. Rather than sensing smoke in the relatively small vicinity of a single sensor, the laser beam approach effectively senses for smoke along a line that can be extended for a considerable distance throughout the monitored area, See Moore, et al., USPN 3,973,852. However, a disadvantage of using such a laser beam approach is that, although it may effectively measure smoke conditions at more points within a monitored area that just those points in the vicinity of a single sensor, it still does not provided feedback on the smoke conditions at all or most of the points within the monitored area. In recent years, multiple practitioners have introduced yet another approach for detecting smoke. Instead of directly measuring the attenuation or scattering of light at the given point in space, their approach uses a digital image that is produced by a TV and identifies the presence of smoke in the viewed area by the effects of the smoke on the video image. In Grech-Cini' s USPN 6,844,818, each pixel of the image is constantly analyzed statistically against the average parameters across the image. The characteristics that are monitored include brightness, color intensity and spatial contrast. Based on the statistical averages, a Bayesian estimator is produced for each pixel and is compared to values from the current image. Once a significant deviation from the estimate occurs over a range of pixels, a determination is made on whether the anomaly is caused by smoke. The disadvantage of this approach is
1 that it has been found to yield a high rate of false alarms that may be caused by
2 changes in the viewed area's lighting conditions or the occurrence of moving
3 objects in the area.
4 In Rizzotti at al.'s USPN 6,937,743, a sequence of images are analyzed to
5 identify changes in the contrast of the image by measuring the attenuation of
6 spatial high frequencies using FFT or FHT. Meanwhile, other practitioners (e.g.,
7 see Toreyin et al., MUSCLE, Malaga, 4-5 Nov 2004) have used a wavelet
8 transform for detecting the of loss in clarity occurring across a sequence of
9 images. io However, it would appear that this entire approach of identifying a image's
! i deterioration of contrast, loss of edges or clarity has fundamental drawbacks.
12 First, its use requires that the content of the image have high contrast with enough
13 details. Second, it is based on an assumption that smoke will be semi-transparent
14 and will gradually attenuate the fine details of the image. In reality, this is assumption is found to be valid in only limited cases. For example, it has been
16 observed that thick smoke, at good illumination conditions and with plain
17 background, can actually increase the high frequency content of an image. is The present inventor has disclosed yet another alternative approach in U.S.
19 Publication No. 2005/0100193, where the light sources within the image are
20 analyzed to detect the smoke-induced effects on diffusion and scattering in the
21 vicinity of these sources. Practical as a supplementary technique, the
22 disadvantage of this approach is its requirement for a permanent light source
23 within the video image and its ability to detect smoke only when smoke particles
24 are in the vicinity of a light source. Similarly, Rattman's USPN 4,614,968
25 suggests using strategically placed multi-contrast markers for an imaging system
26 that provides close-up views of these objects with high contrast and edge contents.
27 However, this approach limits the area of detection to direct lines of sight with the
28 viewed markers.
29 Despite the considerable prior art relating to smoke detectors, there is still
30 a need for smoke detector methods and systems that can more effectively measure
31 smoke conditions throughout the entire volume of a desired monitored area.
32 33
3. OBJECTS AND ADVANTAGES There has been summarized above, rather broadly, the prior art that is related to the present invention in order that the context of the present invention may be better understood and appreciated. In this regard, it is instructive to also consider the objects and advantages of the present invention. It is an object of the present invention to provide apparatus and methods that are effective at detecting smoke within the entire volume of a monitored area. It is another object of the present invention to provide apparatus and methods that are effective at detecting smoke in industrial petrochemical installations. It is an object of the present invention to provide apparatus and methods that can operate within the framework of the ordinary Closed Circuit Television (CCTV) surveillance systems that are used to monitor commercial, outdoor, industrial and residential areas. It is yet another object of the present invention to demonstrate how existing security surveillance equipment may be combined into unique systems which provide the best means to address the detection of smoke in the industrial, commercial and residential installations. It is a further object of the present invention to provide a means for notification of smoky conditions within a monitored area to remote operators who are using closed circuit television to monitor the area. These and other objects and advantages of the present invention will become readily apparent as the invention is better understood by reference to the accompanying summary, drawings and the detailed description that follows.
SUMMARY OF THE INVENTION
Recognizing the need for the development of improved smoke detection systems and methods, the present invention is generally directed to satisfying the needs set forth above and overcoming the disadvantages identified with prior art devices and methods. In accordance with a first embodiment of the present invention, the foregoing need can be satisfied by providing an apparatus for detecting smoke in a monitored area that includes: (a) a means for capturing a series of successive video images of the monitored area as a series of two-dimensional bitmaps having a specified number of pixels, (b) a means for examining this series of bitmaps to identify areas in successive bitmaps of adjacent pixels that have the potential for being used as indicators for the existence of smoke in the monitored area, (c) a means for storing these identified areas from the bitmaps, and (d) a means for analyzing the identified areas so as to identify the temporal changes in these areas which are indicative of the presence of smoke. Alternative embodiments of the present invention can be obtained by adding to this first embodiment any combination of the following elements: (e) a means for analytically characterizing the data contained in each of the identified areas such that the amount of this data is reduced and specifiable information pertaining to these identified areas is preserved so as to allow one to analyze this reduced data to identify the presence of smoke in the monitored area, and (f) a means for generating an alarm when the means for analyzing the identified areas identifies the temporal changes that are indicative of the presence of smoke. Still further embodiments of the present invention can be obtained by configuring these means as follows: (g) the means for examining so that it selects those pixels having an approximate uniform change in intensity, (h) the means for analyzing so that it selects those temporal changes chosen from the group consisting of those temporal changes which indicate an ever-increasing number of pixels in those identified areas of the cluster or those temporal changes which indicate that the increase in the size of the identified areas occurs to a greater extent in the top rather than the bottom part of the identified areas, or (i) the means for analyzing so that it makes use of a neural network.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 shows a block diagram of a preferred embodiment of the smoke detection method and apparatus of the present invention. FIG. 2 shows the algorithm of a preferred embodiment of the smoke detection method and apparatus of the present invention. FIG. 3 illustrates the particular trends that are typical for a growing smoke plume. FIG. 4 illustrates the specific trends that are typical for a walking human. FIG. 5 shows a possible conversion of the growing plume into a 2-D pattern.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Before explaining at least one embodiment of the present invention in detail, it is to be understood that the invention is not limited in its application to the details of construction and to the arrangements of the components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced and carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein are for the purpose of description and should not be regarded as limiting. FIG. 1 shows a block diagram of a preferred embodiment of the present invention. The images of a growing smoke cloud within a specified area are seen to be monitored by a camera 2 that converts the image into a digital representation within the computer memory of a video capturing device 4. This representation takes the form of sequence of digital images which are analyzed by a computer system 6. The analysis of these images consists of: (a) the identification and extraction of the pixel areas (blobs) that are candidates for being representative of smoke clouds (i.e., these blobs have the potential for being used as indicators of smoke in the monitored area), (b) the tracking of these blobs to identify their long term trends, and (c) comparing these trends to identify those which are consistent with those produced by a growing smoke cloud. When such identification is made, a prescribed response system 8 is activated. In every image within a received sequence, the computer system 6 identifies a group of pixels (i.e., blob) that may represent a cloud of smoke by monitoring the changes in the properties of the pixels that comprise the group. Such changes may include, but are not limited to, consistent (across the blob) changes in brightness or intensity. For example, in a simplest case, a cloud of black smoke will make all the pixels within the cloud darker, while white smoke will make all the pixels lighter. If any portion of a blob is spatially overlapping with the blob(s) from the previous image, such blobs are arranged in a tracking cluster that stores all the sequence of overlapping blobs from a series of consecutive frames (i.e., a cluster is only the set of overlapping regions or blobs of the bitmaps; not the entire bitmaps themselves). From my studies of the temporal changes occurring in such
tracking clusters, I have discovered that those of such clusters which are caused by growing smoke clouds exhibit consistent, identifiable signatures in their growth and movement. Using the present invention's discovery that the tracking clusters for various types of physical phenomena of interest (e.g., a smoke cloud) exhibit consistent, identifiable signatures in their growth and movement, one is able to develop an algorithm for analyzing such captured images so as to identify the occurrence of such phenomena of interest. FIG. 2 shows a flow-chart representation of such an algorithm. The analysis scheme of the present invention begins with the identification of a blob of interest. At its simplest form, this step could be achieved by, for example, searching a captured digital image for a group of pixels with consistent brightness changes. Once such a blob has been identified, the analysis scheme checks as to whether this new image's blob should be chronologically associated, on the basis of the likelihood of the existence of smoke in the monitored area, with any of the blobs that are being tracked from earlier images. This is achieved by checking for spatial overlap between this new blob and the database of tracking clusters identified from the prior images. In case of such overlap, the blob is appended to the appropriate tracking cluster. If no such overlap exists, this blob becomes the first of what may be a sequence of blobs that will make up yet another cluster. Clusters that have no new blobs added to them are assumed to no longer be active and are purged from the database. Each updated tracking cluster is analyzed to identify whether the addition of this new blob allows one to identify in such clusters any of the special patterns or signatures of growth or movement that can be associated with a particular type of physical phenomena (e.g., growth and movement of a smoke cloud). When such a signature has been positively identified, a prescribed system response can be activated. Meanwhile, this analysis scheme proceeds to analyze the next blob. A representative example of the identifiable signatures in the growth and movement of the blobs within the clusters for growing smoke clouds is shown in FIG. 3. The contours of the blobs from consecutive video frames 1, 2 and 3 are seen to have spatial overlay.
The changes in the growth, movement or shape of representative smoke cloud blobs can be caused by several factors, including: (a) constant supply of smoke from a source, (b) diffusion of the smoke, and (c) air movements within the monitored space. Examples of the blob movements or size changes that can be used to identify such smoke clouds include: (i) blob boundary expansions as more smoke is generated, (U) blob boundary movement and expansion towards the top of the blob due to the hotter gases that comprise the smoke cloud rising in the surrounding, cooler ambient air, and (iii) some area of blob overlay between consecutive images. Such blob movements can be easily distinguished from those of other types of physical phenomena or the movements of various types of objects within the view of a camera that captures a temporal record of their digital images. For example, FIG. 4 shows the evolution of the blob that is caused by a human moving across the field of view of a camera. Technical barriers to the implementation of various embodiments of the present invention include the need to accommodate for the "fuzziness" which is usually seen in actual video data. One means of addressing this problem is the use of a feed-forward, three layer, neural network (those knowledgeable in such technology will recognize that similar results may be achieved by utilizing different neural network layouts). Meanwhile, the magnitude of necessary calculations and the complexity of the underlying network can be simplified by scaling the entire rectangular area of an image tracking cluster into fixed-size pattern of size, for example, 12x12. For illustrative purposes, it proves useful to show one of the conceivably many possible data reduction schemes that can be used by the present invention for expressing the information pertaining to a typical smoke cluster in a form that will make the further analysis of such data more tractable in terms of allowing one to more easily and with fewer numerical computations identify in such data the signatures or trends that are indicative of smoke in a monitored area. FIG. 5 shows a three dimensional (time plus a two dimensional plane) space for the representation and analysis of such clusters. The vertical axis of this space is defined to be the time since a desired sequence of video images began to be accumulated, while its horizontal, X-Y, plane represents the spatial extent of a video image.
At any given instant, the digital image of a smoke cloud is assumed to be a solid, two dimensional figure (i.e., it has no interior voids; it has a continuous enclosing contour or boundary) such as those represented by one of the slices that are stacked one above another in the temporal plane of FIG. 5. The lateral extent of these slices is seen to be increasing in time (i.e., the sizes of the images taken at instants that are a considerable time after the initiation of the smoke cloud are larger than those taken when the smoke cloud had just begun to form). In the lower, left-hand side of FIG. 5 is shown a two-dimensional, normalized (12x12) overlay of these slices (i.e., a footprint) in which each cell of this footprint represents the relative (to age of the cluster) time span that the particular cell has been covered by the smoke cloud, and the phase (time of initiation relative to age of the cluster) when the smoke cloud first covered the cell in question. The analytical representation of the information contained in FIG. 3 can be further simplified, for example, by allowing each cell in this footprint to be represented by a 2-d vector in which the phase, α, of this vector is defined so as to provide information on the time when the pixel was "energized" or at what exact moment pixel has been assigned smoke status (e.g., α ranges from 0 to π/2, with π/2 being defined as the youngest pixels, or those that have been identified at the greatest time since the initiation of the smoke cloud was recognized). The length, ξ, of this 2-d vector may be used to encode, for example, the "age of cell," which is the time duration that it has been covered by the cloud. In such an analytical representation, the typical smoke cloud yields a cluster that has its longest vector with close to zero phase at the location of the smoke's source, while shorter and similarly higher phased vectors will be at the peripheral locations of the smoke cloud. Once all necessary information is encoded into a chosen data reduction scheme, it is matter of training of the previously mentioned, neural network to successfully identify the trends and signatures of the various phenomena that are represented in the captured images. Those familiar with such data analysis techniques should realize that neural networks are by no means the only technique that is available to identify the signatures of various phenomena in such sequences of visual images. Various alternative approaches that are considered to come within the scope of the present
invention for the identification of a smoke cloud include: (a) trend analysis of a smoke cloud's contour area (e.g., an ever-increasing number of pixels in those identified areas of a cluster, or the identification of cloud's contour area changes as a positive linear trend over a prescribed period of time with the computed least squares value for said trend being below a prescribed level, or the identification of cloud's contour area changes by a polynomial least square approximation with a consistent positive first derivative exceeding a prescribed value over the a prescribed period and with the least square value being below a prescribed level), (b) movement analysis in the top and bottom segments of a cloud contour (e.g., the increase in the size of the identified areas occurs to a greater extent in the top rather than the bottom part of the identified areas, where the top and bottom parts are defined with respect to the direction of influence of gravity), and (c) various other, logical, rule-based, predicative approaches. In a preferred embodiment, the present invention is implemented on an Intel® Pentium™ IV class processor platform under a Windows operating system. This processing system is capable of simultaneously handling eight video channels and can also provide the necessary alarm notifications over EP networks. The algorithms of the present invention have been implemented in C++ programming language using the Microsoft Visual Studio .NET environment. A feed forward, three-layer neural network (294/20/1) configuration is used for pattern matching. Neural network training for smoke cloud identification was done on over 10,000 samples from which more than 95% are nuisances. In concluding this disclosure, it should be noted that the foregoing is considered as illustrative only of the principles of the present invention. Further, since numerous modifications and changes will readily occur to those skilled in the art. it is not desired to limit the invention to the exact methods described herein, and accordingly, all suitable modifications and equivalents may be resorted to which fall within the scope of the invention as hereinafter forth in the claims to it.
Claims
i CLAIMS
2
3 I claim:
4 1. An apparatus for detecting smoke in a monitored area, said apparatus
5 comprising:
6 a means for capturing a series of successive video images of said
7 monitored area in the form of a series of two-dimensional bitmaps having a
8 specified number of pixels,
9 a means for examining said series of bitmaps to identify areas in
10 successive bitmaps of adjacent pixels that have the potential for being used as
11 indicators for the existence of smoke in said monitored area,
12 wherein said resulting identified areas from said series of bitmap forming a B cluster of identified areas,
14 a means for storing said cluster of identified areas from said bitmaps, and is a means for analyzing said cluster of identified areas so as to identify the
16 temporal changes in said areas which are indicative of the presence of smoke in
17 said series of successive video images. is 2. The apparatus as recited in Claim 1 , further comprising:
19 a means for characterizing, prior to said analysis step, the data contained in
20 each of said identified areas such that the amount of said data is reduced and
21 specifiable information pertaining to said identified areas is preserved so as to
22 allow one to analyze said reduced data to identify the presence of smoke in said
23 monitored area, and
24 wherein said means for analyzing said identified areas is configured so as
25 to operate on said reduced data.
26 3. The apparatus as recited in Claim lor Claim 2, further comprising:
27 a means for generating an alarm when said means for analyzing said
28 identified areas identifies said temporal changes that are indicative of the presence
29 of smoke.
30 4. The apparatus as recited in any one of Claims 1 to 3, wherein:
31 said means for examining said series of bitmaps, to identify areas in
32 successive bitmaps of adjacent pixels that have the potential for being used as
33 indicators for the existence of smoke in said monitored area, is configured so that
said means accomplishes said identification by selecting those pixels having an approximate uniform change in intensity. 5. The apparatus as recited in any one of Claims 1 to 4, wherein: said means for analyzing said cluster of identified areas is configured so that said means accomplishes said analysis by selecting said temporal changes chosen from the group consisting of those temporal changes which indicate an ever-increasing number of pixels in those identified areas of said cluster or those temporal changes which indicate that the increase in the size of said identified areas occurs to a greater extent in the top rather than the bottom part of said identified areas, and said top and bottom portion of said identified areas being defined with respect to the direction of influence of gravity within said monitored area. 6. The apparatus as recited in any one of Claims 1 to 5, wherein: said means for analyzing said cluster of identified areas is configured so that said means accomplishes said analysis by making use of a neural network. 7. A method for detecting smoke in a monitored area, said method comprising the steps of: capturing a series of successive video images of said monitored area in the form of a series of two-dimensional bitmaps having a specified number of pixels, examining said series of bitmaps to identify areas in successive bitmaps of adjacent pixels that have the potential for being used as indicators for the existence of smoke in said monitored area, wherein said resulting identified areas from said series of bitmap forming a cluster of identified areas, storing said cluster of identified areas from said bitmaps, and analyzing said cluster of identified areas so as to identify the temporal changes in said areas which are indicative of the presence of smoke in said series of successive video images. 8. The method as recited in Claim 7, further comprising the step of: characterizing, prior to said analysis step, the data contained in each of said identified areas such that the amount of said data is reduced and specifiable information pertaining to said identified areas is preserved so as to allow one to analyze said reduced data to identify the presence of smoke in said monitored area, and
wherein said step of analyzing said identified areas is performed on said reduced data. 9. The method as recited in any one of Claims 7 or 8, wherein said examining step involves accomplishing said identification by selecting those pixels having an approximate uniform change in intensity, said analyzing step involves selecting temporal changes chosen from the group consisting of those temporal changes which indicate an ever-increasing number of pixels in those identified areas of said cluster or those temporal changes which indicate that the increase in the size of said identified areas occurs to a greater extent in the top rather than the bottom part of said identified areas, and wherein top and bottom portion of said identified areas being defined with respect to the direction of influence of gravity within said monitored area. 10. The method as recited in any one of Claims 7 to 9, wherein said step of analyzing said cluster of identified areas involves making use of a neural network.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP07750488.4A EP1994502B2 (en) | 2006-02-13 | 2007-02-13 | Smoke detection method and apparatus |
DE602007009649T DE602007009649D1 (en) | 2006-02-13 | 2007-02-13 | METHOD AND DEVICE FOR SMOKE DETECTION |
AT07750488T ATE484039T1 (en) | 2006-02-13 | 2007-02-13 | METHOD AND DEVICE FOR SMOKE DETECTION |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US11/352,721 | 2006-02-13 | ||
US11/352,721 US7769204B2 (en) | 2006-02-13 | 2006-02-13 | Smoke detection method and apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2007095160A2 true WO2007095160A2 (en) | 2007-08-23 |
WO2007095160A3 WO2007095160A3 (en) | 2008-11-20 |
Family
ID=38255627
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US2007/003655 WO2007095160A2 (en) | 2006-02-13 | 2007-02-13 | Smoke detection method and apparatus |
Country Status (5)
Country | Link |
---|---|
US (1) | US7769204B2 (en) |
EP (1) | EP1994502B2 (en) |
AT (1) | ATE484039T1 (en) |
DE (1) | DE602007009649D1 (en) |
WO (1) | WO2007095160A2 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101726357A (en) * | 2008-10-14 | 2010-06-09 | 能美防灾株式会社 | Smoke detecting apparatus |
Families Citing this family (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5108893B2 (en) * | 2006-10-27 | 2012-12-26 | トムソン ライセンシング | System and method for restoring a 3D particle system from a 2D image |
US8229227B2 (en) * | 2007-06-18 | 2012-07-24 | Zeitera, Llc | Methods and apparatus for providing a scalable identification of digital video sequences |
US8462980B2 (en) * | 2008-05-08 | 2013-06-11 | Utc Fire & Security | System and method for video detection of smoke and flame |
TWI385602B (en) * | 2008-10-17 | 2013-02-11 | Ind Tech Res Inst | Smoke detecting method and system |
CN101751745B (en) * | 2008-12-16 | 2013-02-20 | 深圳富泰宏精密工业有限公司 | Portable electronic device with smoke detection function and smoke detection method |
US9140646B2 (en) | 2012-04-29 | 2015-09-22 | Valor Fire Safety, Llc | Smoke detector with external sampling volume using two different wavelengths and ambient light detection for measurement correction |
US8947243B2 (en) | 2012-04-29 | 2015-02-03 | Valor Fire Safety, Llc | Smoke detector with external sampling volume and utilizing internally reflected light |
US8907802B2 (en) | 2012-04-29 | 2014-12-09 | Valor Fire Safety, Llc | Smoke detector with external sampling volume and ambient light rejection |
JP6139977B2 (en) * | 2013-05-22 | 2017-05-31 | ホーチキ株式会社 | Fire detection device and fire detection method |
JP6139982B2 (en) * | 2013-05-24 | 2017-05-31 | ホーチキ株式会社 | Fire detection device and fire detection method |
DE102013017395B3 (en) | 2013-10-19 | 2014-12-11 | IQ Wireless Entwicklungsges. für Systeme und Technologien der Telekommunikation mbH | Method and device for automated early forest fire detection by means of optical detection of clouds of smoke |
EP3063747B1 (en) | 2013-10-30 | 2019-07-24 | Valor Fire Safety, LLC | Smoke detector with external sampling volume and ambient light rejection |
JP6345931B2 (en) * | 2013-12-27 | 2018-06-20 | ホーチキ株式会社 | Fire detection system and fire detection method |
JP6317116B2 (en) * | 2014-01-24 | 2018-04-25 | ホーチキ株式会社 | Fire detection system and fire detection method |
GB2528195B (en) * | 2015-09-17 | 2017-03-08 | Micropack (Engineering) Ltd | Flame detection in an image sequence |
JP6718646B2 (en) * | 2016-04-15 | 2020-07-08 | ホーチキ株式会社 | Fire detection device and fire detection method |
WO2018116966A1 (en) * | 2016-12-21 | 2018-06-28 | ホーチキ株式会社 | Fire monitoring system |
US11188762B2 (en) * | 2017-05-10 | 2021-11-30 | Konica Minolta, Inc. | Moving body risk assessment device, and moving body monitoring system |
CN108334902A (en) * | 2018-02-02 | 2018-07-27 | 北京华纵科技有限公司 | A kind of track train equipment room smog fireproof monitoring method based on deep learning |
GB2592463B (en) | 2019-06-27 | 2023-05-17 | Carrier Corp | Spatial and temporal pattern analysis for integrated smoke detection and localization |
EP4000052A1 (en) | 2019-07-18 | 2022-05-25 | Carrier Corporation | Flame detection device and method |
CN111652184B (en) * | 2020-06-19 | 2023-06-20 | 成都通甲优博科技有限责任公司 | Smoke identification method, device, storage medium and data processing equipment |
CN112859652B (en) * | 2021-01-20 | 2022-03-29 | 青岛海尔空调器有限总公司 | Indoor smoking monitoring method and monitoring device |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050100193A1 (en) | 2003-11-07 | 2005-05-12 | Axonx, Llc | Smoke detection method and apparatus |
US6937743B2 (en) | 2001-02-26 | 2005-08-30 | Securiton, AG | Process and device for detecting fires based on image analysis |
Family Cites Families (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3973852A (en) | 1974-08-30 | 1976-08-10 | The Dow Chemical Company | Method and apparatus for measuring particulate concentration in the atmosphere |
US4170264A (en) | 1977-07-27 | 1979-10-09 | Gibson Motor And Machine Service, Inc. | Pump and roll, vehicle with an elevatable water tower |
US4875526A (en) | 1988-12-09 | 1989-10-24 | Latino Vincent P | Rough terrain, large water volume, track driven firefighting apparatus and method |
US5153722A (en) | 1991-01-14 | 1992-10-06 | Donmar Ltd. | Fire detection system |
US5237308A (en) * | 1991-02-18 | 1993-08-17 | Fujitsu Limited | Supervisory system using visible ray or infrared ray |
FR2696939B1 (en) * | 1992-10-16 | 1995-01-06 | Bertin & Cie | Method and device for rapid automatic detection of forest fires. |
CH686913A5 (en) | 1993-11-22 | 1996-07-31 | Cerberus Ag | Arrangement for early detection of fires. |
US5719557A (en) | 1994-05-19 | 1998-02-17 | Digital Security Controls Ltd. | Photoelectric smoke detector |
US5832187A (en) * | 1995-11-03 | 1998-11-03 | Lemelson Medical, Education & Research Foundation, L.P. | Fire detection systems and methods |
AUPN965996A0 (en) | 1996-05-03 | 1996-05-30 | Vision Products Pty Ltd | The detection of airborne pollutants |
JP3481397B2 (en) | 1996-07-29 | 2003-12-22 | 能美防災株式会社 | Fire detector |
US5850182A (en) | 1997-01-07 | 1998-12-15 | Detector Electronics Corporation | Dual wavelength fire detection method and apparatus |
US5995008A (en) | 1997-05-07 | 1999-11-30 | Detector Electronics Corporation | Fire detection method and apparatus using overlapping spectral bands |
US5838242A (en) | 1997-10-10 | 1998-11-17 | Whittaker Corporation | Fire detection system using modulation ratiometrics |
DE19840873A1 (en) † | 1998-09-01 | 2000-03-09 | Deutsch Zentr Luft & Raumfahrt | Method and device for automatic forest fire detection |
GB9822956D0 (en) * | 1998-10-20 | 1998-12-16 | Vsd Limited | Smoke detection |
GB9922761D0 (en) * | 1999-09-27 | 1999-11-24 | Sentec Ltd | Fire detection algorithm |
ATE340395T1 (en) * | 2000-02-07 | 2006-10-15 | Vsd Ltd | SMOKE AND FLAME DETECTION |
SE524332C2 (en) * | 2000-03-20 | 2004-07-27 | Karl-Erik Morander | System and method for optical monitoring of a volume |
US6184792B1 (en) | 2000-04-19 | 2001-02-06 | George Privalov | Early fire detection method and apparatus |
US7280696B2 (en) * | 2002-05-20 | 2007-10-09 | Simmonds Precision Products, Inc. | Video detection/verification system |
AU2003902319A0 (en) * | 2003-05-14 | 2003-05-29 | Garrett Thermal Systems Limited | Laser video detector |
DE102004056958B3 (en) † | 2004-11-22 | 2006-08-10 | IQ wireless GmbH, Entwicklungsgesellschaft für Systeme und Technologien der Telekommunikation | Surveillance of territories for detection of forest and wildfires |
-
2006
- 2006-02-13 US US11/352,721 patent/US7769204B2/en active Active
-
2007
- 2007-02-13 AT AT07750488T patent/ATE484039T1/en not_active IP Right Cessation
- 2007-02-13 EP EP07750488.4A patent/EP1994502B2/en active Active
- 2007-02-13 DE DE602007009649T patent/DE602007009649D1/en active Active
- 2007-02-13 WO PCT/US2007/003655 patent/WO2007095160A2/en active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6937743B2 (en) | 2001-02-26 | 2005-08-30 | Securiton, AG | Process and device for detecting fires based on image analysis |
US20050100193A1 (en) | 2003-11-07 | 2005-05-12 | Axonx, Llc | Smoke detection method and apparatus |
Non-Patent Citations (2)
Title |
---|
GEORGE PRIVALOV; ALEXANDER SHAYKHUTDINOV'S: "Machine vision delivers fire protection", VISION SYSTEMS DESIGN PENNWELL CORP USA, vol. 9, no. 3, March 2004 (2004-03-01), pages 31 - 35 |
TOREYIN ET AL., MUSCLE, 4 November 2004 (2004-11-04) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101726357A (en) * | 2008-10-14 | 2010-06-09 | 能美防灾株式会社 | Smoke detecting apparatus |
Also Published As
Publication number | Publication date |
---|---|
US7769204B2 (en) | 2010-08-03 |
DE602007009649D1 (en) | 2010-11-18 |
EP1994502A2 (en) | 2008-11-26 |
US20070188336A1 (en) | 2007-08-16 |
EP1994502B1 (en) | 2010-10-06 |
ATE484039T1 (en) | 2010-10-15 |
WO2007095160A3 (en) | 2008-11-20 |
EP1994502B2 (en) | 2019-09-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7769204B2 (en) | Smoke detection method and apparatus | |
US7859419B2 (en) | Smoke detecting method and device | |
EP1687784B1 (en) | Smoke detection method and apparatus | |
CA2795896C (en) | Method and system for security system tampering detection | |
JP4705090B2 (en) | Smoke sensing device and method | |
AU2017359142B2 (en) | Systems and methods for detecting flying animals | |
CA2275893C (en) | Low false alarm rate video security system using object classification | |
WO2000034913A1 (en) | Method for rejection of flickering lights in an imaging system | |
EP2000998B1 (en) | Flame detecting method and device | |
JP2012033152A (en) | System and method for classifying moving object upon monitoring via video | |
CN114664048A (en) | Fire monitoring and fire early warning method based on satellite remote sensing monitoring | |
CN112643719A (en) | Tunnel security detection method and system based on inspection robot | |
Lai et al. | Robust little flame detection on real-time video surveillance system | |
Dimitropoulos et al. | Flame detection for video-based early fire warning for the protection of cultural heritage | |
CN115880598B (en) | Ground image detection method and related device based on unmanned aerial vehicle | |
JP2005284652A (en) | Video monitoring method and apparatus using moving vector | |
Lee et al. | Fire detection using color and motion models | |
JP4617905B2 (en) | Monitoring device and method, recording medium, and program | |
JP5309069B2 (en) | Smoke detector | |
Dulski et al. | Data fusion used in multispectral system for critical protection | |
Bistrović et al. | Computer Vision Application for Early Stage Smoke Detection on Ships | |
Ciurapiński et al. | Data fusion concept in multispectral system for perimeter protection of stationary and moving objects | |
Craciunescu et al. | Overview of image processing tools to extract physical information from JET videos | |
Wenger | A framework based on Gaussian mixture models and Kalman filters for the segmentation and tracking of anomalous events in shipboard video | |
CN115019463A (en) | Water area supervision system based on artificial intelligence technology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2007750488 Country of ref document: EP |