World Applied Sciences Journal 5 (1): 000-000, 2008 ISSN 1818-4952
© IDOSI Publications, 2008
Corresponding Author: Dr. Moein Shakeri, Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
1
Fuzzy-Cellular Background Subtraction Technique for Urban Traffic Applications
Moein Shakeri, Hossein Deldari
Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
Abstract: Computational structure of cellular automata has attracted researchers and vastly been used in various fields of science. Cellular automata perform complex computations with a high degree of efficiency and robustness; they are especially suitable for modeling natural systems that can be described as massive collections of simple objects interacting locally with each other, such as motion detection in image processing. On the other hand, extraction of moving objects from an image sequence is a fundamental problem in dynamic image analysis. A common method for real-time segmentation of moving regions in image sequences, involves “background subtraction. Nowadays background modeling and subtraction algorithms are commonly used in real-time urban traffic applications for detecting and tracking vehicles and monitoring streets. In this paper by the use of cellular automata, a novel fuzzy approach for background subtraction with a particular interest to the problem of vehicle detection is presented. Our experimental results demonstrate that fuzzy-cellular system is much more efficient, robust and accurate than classical approaches.
mo_sh88@stu-mail.um.ac.ir, hd@um.ac.ir
Key words: Fuzzy background subtraction • fuzzy-cellular background modeling • cellular
automata • vehicle detection
INTRODUCTION
Identifying moving objects from a video sequence is a fundamental and critical task in video surveillance, traffic monitoring and analysis, human detection and tracking, and gesture recognition in human-machine interface. A common approach to identifying the moving objects is background subtraction, where each video frame is compared against a reference or background model. Pixels in the current frame that deviate significantly from the background are considered to be moving objects. These “foreground" pixels are further processed for object localization and tracking. Since background subtraction is often the first step in many computer vision applications, it is important that the extracted foreground pixels accurately correspond to the moving objects of interest. There are several problems that a good background subtraction algorithm must solve correctly. Consider a video sequence from a stationary camera overlooking a traffic intersection. As it is an outdoor environment, a background subtraction algorithm should adapt to various levels of illumination at different times of the day and handle adverse weather condition such as fog or snow that modifies the background. Changing shadow, cast by moving objects, should be removed so that consistent features can be extracted from the
objects in subsequent processing. The complex traffic flow at the intersection also poses challenges to a background subtraction algorithm. The vehicles move at a normal speed when the light is green, but come to a stop when it turns red. The vehicles then remain stationary until the light turns green again. A good background subtraction algorithm must handle the moving objects that first merge into the background and then become foreground at a later time. In addition, to accommodate the real-time needs of many applications, a background subtraction algorithm must be computationally inexpensive and have low memory requirements, while still being able to accurately identify moving objects in the videos [1].
According to importance of real-time computations in the surveillance systems, improvement the efficiency of simple background subtraction methods is so significant. To this end, in this paper we propose a novel fuzzy method based on computational model of cellular automata for background subtraction and moving object detection. Our experimental results demonstrate that the fuzzy-cellular system is much more efficient, robust and accurate than classical approaches.
