Image analysis – Pattern recognition – Feature extraction
Patent
1995-05-10
1997-01-07
Couso, Jose L.
Image analysis
Pattern recognition
Feature extraction
382203, G06K 900
Patent
active
055925677
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
In the processing of sequences of digital images from highway traffic scenes for the purposes of detection, tracking and classification of vehicles, there is often the disturbing effect that the vehicles throw a shadow on the highway in the event of there being direct sunshine, the shadow not being able to be separated in a simple way from the vehicle. As a result, a classification of vehicles according to their geometrical dimensions, in particular according to their width, is often made more difficult, if not impossible.
Previous approaches to solving this problem fall back on vehicle models (G. D. Sullivan, K. D. Baker: Model-based vision: using cues to select hypotheses. SPIE Vol. 654 Automatic Optical Inspection (1986), pp. 272-277), which are brought into coincidence with the image (matching), as a result of which separating the shadow is not necessary. This approach has the disadvantage, however, that it is very computation-intensive and that, given the present state of development of image processing hardware, an implementation of such methods in real time within a useful cost frame does not appear possible.
SUMMARY OF THE INVENTION
The present invention is therefore based on the object of specifying a method for detecting and separating the shadow of moving objects in a sequence of digital images, which is suitable for application in the above-mentioned image processing methods and which can be implemented in real time on the basis of cost-effective hardware modules. According to the invention, this object is achieved by a method for detecting and separating the shadow of moving objects in a sequence of digital images having the following steps.
In this method, a sequence of background images is first determined from a sequence of images, the sequence of background images containing only the non-moving image background of the sequence of images but not the moving objects. Object edges are then extracted from the sequence of images, in that, for each image of the sequence of images and for the respective corresponding background image of the sequence of background images, in each case an edge gray-value image is determined. By means of vertical or horizontal projection of the edge gray-value images, corresponding vertical or horizontal edge histograms are determined in each case for the edge gray-value images. Edges which do not belong to moving objects or shadows of moving objects are eliminated, in that the edge histograms of a background image are subtracted from the corresponding edge histograms of the image, belonging to the same instant, of the sequence of images and in that the differential edge histograms formed in this manner are subjected to a threshold value discrimination, by which means those edges are eliminated which do not correspond to the actual moving objects but rather to a shadow of a moving object.
To carry out this method, an essentially smaller computational cost is necessary than in the case of matching of object models to image regions.
In a preferred embodiment of the invention, the edge gray-value images are put into binary form before the determination of the edge histograms. By means of this measure, the computational cost in the processing of the edge histograms can be further reduced.
A further reduction of the computational cost becomes possible if, before carrying out the steps of extracting object edgesn1, those image regions are determined which correspond to moving objects - if appropriate including their shadow - and in which the remaining steps are applied only to these image regions.
In a further preferred embodiment of the method according to the invention, edge histograms which belong to one and the same moving object are combined into a time sequence of edge histograms and the elimination of shadow edges is improved by means of an identification of edges corresponding to each other.
In a further preferred embodiment of the method, the elimination of the shadow edges is improved further in that preknowledge about the lateral
REFERENCES:
patent: 3760377 (1973-09-01), Attridge et al.
patent: 4433325 (1984-02-01), Tanaka et al.
patent: 4954962 (1990-09-01), Evans, Jr. et al.
patent: 5099324 (1992-03-01), Abe
patent: 5293235 (1994-03-01), Guede et al.
patent: 5410346 (1995-04-01), Saneyoshi et al.
patent: 5410418 (1995-04-01), Yonezawa
patent: 5428723 (1995-06-01), Ainscow et al.
"Measurement of Two-Dimensional Movement of Traffic by Image Processing", Kobatake et al., IEEE Proceedings: ICASSP 87 International Conference on Acoustics, Speech & Signal Processing, Apr. 6-9, 1987, pp. 614-617.
"Detection Et Suivi D'Objets Par Analyse De Sequences D'Images", Pineau et al., Treizieme Colloque Sur Le Traitement Du Signal Et Des Images Juan-Les-Pins, Sep. 16-20, 1991, pp. 1125-1128.
"Optical Sensing and Size Discrimination of Moving Vehicles Using Photocell Array and Threshold Devices", Tasuku Takagi, IEEE Transactions on Instrumentation and Measurement, vol. 25, No. 1, Mar. 1976, pp. 52-55.
"Moving Object Segmentation Based on Adaptive Reference Images", Karmann et al., Proceedings of Eusipco, Barcelona, Spain, Sep. 18-21, 1990, pp. 951-954.
"A Video-Based System for Extracting Traffic Flow Parameters", W. Feiten et al., Proceedings from 13th DAGM Symposium, Munich, Oct. 9-11, 1991, pp. 507-514.
Couso Jose L.
Siemens Aktiengesellschaft
Tadayon Bijan
LandOfFree
Method for detecting and separating the shadow of moving objects does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for detecting and separating the shadow of moving objects, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for detecting and separating the shadow of moving objects will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1771785