Motion detection in color image sequence and shadow elimination

Most of the researches are concentrated on motion detection in gray value image sequences and the methods for motion detection are based on background subtraction or on temporal gray value derivatives. The methods based on background subtraction, including auto-adaptive ones, meet difficulties in presence of illumination changes and of slowly moving objects and need to be re-initialized from time to time. The methods based on temporal derivatives are in general sensible to noise. Color images containing much richer information than the gray value ones, it would be interesting to use them to better detect moving objects. In this paper, we address the problem of motion detection in color image sequences and the problems of illumination changes and shadow elimination. Our motion detection method is based on fuzzy segmentation of the color difference image in help of non-symmetrical π membership functions. The elimination of false moving objects detected due to illumination change is realized by combining the background subtraction method with the temporal derivative method and motion continuity. Shadows are removed by comparing the color of mobile pixels detected in the current frame with that in the precedent frame in HSL color space. Experimental results are reported.

[1]  Paola Mello,et al.  Image analysis and rule-based reasoning for a traffic monitoring system , 1999, Proceedings 199 IEEE/IEEJ/JSAI International Conference on Intelligent Transportation Systems (Cat. No.99TH8383).

[2]  D. Koller,et al.  Towards robust automatic traffic scene analysis in real-time , 1994, Proceedings of 1994 33rd IEEE Conference on Decision and Control.

[3]  Jun Shen,et al.  Multi-edge detection by isotropical 2-D ISEF cascade , 1995, Pattern Recognit..

[4]  Mitra Tzi - ker ChiuehComputer S ien e DepartmentState.  Three-Dimensional Computer , 2000 .

[5]  Jun Shen,et al.  Towards the unification of band-limited derivative operators for edge detection , 1993, Signal Process..

[6]  Hironobu Fujiyoshi,et al.  Moving target classification and tracking from real-time video , 1998, Proceedings Fourth IEEE Workshop on Applications of Computer Vision. WACV'98 (Cat. No.98EX201).

[7]  Paul L. Rosin.  Unimodal thresholding , 2001, Pattern Recognit..

[8]  Jun Shen.  On Multi-Edge Detection , 1996, CVGIP Graph. Model. Image Process..

[9]  Robert C. Bolles,et al.  Background modeling for segmentation of video-rate stereo sequences , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[10]  Trevor Darrell,et al.  Background estimation and removal based on range and color , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[11]  Wei Shen,et al.  Fuzzy neural nets with non-symmetric pi membership functions and applications in signal processing and image analysis , 2000, Signal Processing.

[12]  Rita Cucchiara,et al.  Improving shadow suppression in moving object detection with HSV color information , 2001, ITSC 2001. 2001 IEEE Intelligent Transportation Systems. Proceedings (Cat. No.01TH8585).

[13]  Olaf Munkelt,et al.  Adaptive Background Estimation and Foreground Detection using Kalman-Filtering , 1995 .

[14]  W. Eric L. Grimson,et al.  Adaptive background mixture models for real-time tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).