Communications: electrical – Condition responsive indicating system – Specific condition
Reexamination Certificate
2001-09-28
2004-11-02
Trieu, Van T. (Department: 2632)
Communications: electrical
Condition responsive indicating system
Specific condition
C340S604000, C340S605000
Reexamination Certificate
active
06812846
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to spill detectors that employ image-processing techniques to the detection of spills, particularly spills that pose a risk to building occupants such as those that occur when containers are accidentally broken in high traffic areas such as supermarkets or warehouses.
2. Background
Spills of food and other materials in public venues, such as stores, present a serious health hazard. Yet spills occur with some frequency in public places, such as supermarkets, office buildings, commuting stations, etc. The longer a spill remains in a public place, the greater the danger of a serious mishap, such as a slip by an occupant. The delay between the time a spill is detected, reported, and cleaned-up, may be lengthy in most circumstances. Providing sufficient manpower to detect spills quickly is expensive.
Therefore, there is a need in the art for mechanisms for detecting spills that do not rely on human observers and which can notify cleanup personnel quickly so that appropriate action can be taken.
SUMMARY OF THE INVENTION
A spill detector uses optical image-processing hardware and software to identify spills. In an embodiment, one or more cameras capture images and/or video sequences, which are then digitally processed to determine if a spill is evident in the captured scene(s). In embodiments, still images are regularly captured and averaged to create a fixed background image to which a current image may be compared. One or more current images may be segmented and the segments further processed to identify candidate segments identifiable as spills. Segmentation may be performed using known techniques in image analysis and object recognition fields such as region-growing, edge-connecting, or other techniques used for distinguishing fields with homogeneous characteristics such as color, luminance, discrete cosine transform (DCT) profiles, etc. The image(s) is(are) further processed to distinguish candidate spill segments from other segments.
Segments identifiable with the normal stationary background and segments identifiable with non-spill foreground objects, such as people or cars, are preferably removed from the set of candidate segments identifiable with a spill. For example, segments that do not change over time—background segments—may be subtracted, using known object-recognition techniques from the field of image processing aimed at object-recognition. Background subtraction may be used to remove portions of the image before segmentation as well. Segments that move from one image to the next in a current sequence may be identified as non-spill foreground segments. Stereo or multiple camera views may be used to determine the elevation of segments and portions thereof above the floor. Segments that correspond to objects above a range of heights, or ones that are not substantially planar may be removed from the candidate segments.
Candidate segments may be further characterized by luminance and/or chrominance histograms (or “maps”) which may be compared to signatures associated with known types of spills. For example, thin runny liquids spills have a high proportion of specular albido and cereal spills a high proportion of diffuse albido and thus correspond to particular luminance maps that may be substantially different from the background or typical non-spill foreground objects.
Where respective cameras form multiple images from different perspectives, the three-dimensional surface shape of segments may be tested to determine if they are planar, as are most hazardous spills such as runny liquids. Thus, a segment in one camera's image may be warped to fit the expected projection of a flat surface, located at the level of the floor, in the other camera's field of view. The segments of the other camera and the predicted segment may then be compared in terms of boundaries, luminance and chrominance maps, and other characteristics to determine if the segments are the same flat floor-level object.
Yet another filtering technique may be to identify occupants and observe avoidance behavior. In other words, occupants would be expected to avoid spills they encounter. Foreground objects may be identified as travelling occupants using known techniques for example as used for occupant-counting. Candidate spill objects may be tested by evaluating observed movement, such as movement that is apparently to avoid the candidate object. Other sources of information may also be used to discount the measure of reliability of the spill detection. For example, the sound of glass breaking or a crashing sound associated with a spill may be recognized by an audio signature and correlated with the above-discussed image/video features. Infrared imaging may be used to identify wet surfaces of spills that are cooled below the ambient temperature by evaporation. Chemical sniffers may detect the presence of certain chemicals in the air, as a result of a spill, to augment the reliability of detection.
The invention will be described in connection with certain preferred embodiments, with reference to the following illustrative figures so that it may be more fully understood. With reference to the figures, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the preferred embodiments of the present invention only, and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the invention. In this regard, no attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the invention may be embodied in practice.
REFERENCES:
patent: 3899213 (1975-08-01), Fantasia et al.
patent: 5257085 (1993-10-01), Ulich et al.
patent: 5281826 (1994-01-01), Ivancic et al.
patent: 6008492 (1999-12-01), Slater et al.
patent: 6084510 (2000-07-01), Lemelson et al.
patent: 7-73384 (1995-03-01), None
Gutta Srinivas
Philomin Vasanth
Koninklijke Philips Electronics , N.V.
Thorne Gregory L.
Trieu Van T.
LandOfFree
Spill detector based on machine-imaging does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Spill detector based on machine-imaging, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spill detector based on machine-imaging will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3316485