Real Time Segmentation Algorithm for Complex Outdoor Conditions

Abstract

Automatic detection of moving objects in all weather condition is a critical task in many vision based safety applications. Moving object detection and segmentation in an outdoor environment, particularly under non ideal weather conditions and unfavorable luminance condition is still an active area in research. In the past, full-search sum of absolute difference algorithm is used to detect and segment the moving object under sudden changes in illumination, snowfall, fog etc.This algorithm is computationally expensive and inefficient for human and night time detection. So we propose a robust technique to detect and segment the human and other moving objects of interest in both day and night environment using CSCA (contrast and skin color analysis) algorithm, which not only reduces the false motion but improves the computational efficiency as well. Experimental result shows that our proposed CSCA algorithm is efficient in motion detection and segmentation in both day and night complex environment conditions and is suitable for real time implementation.