Moving Vehicle Detection in Dynamic Background from Airborne Monocular Camera

Aerial imagery sensors can provide sufficient resolution to sense vehicle locations and movements across broader spatial and temporal scales. In this paper, an approach for collecting and analyzing aerial imagery is given. This paper presents a method to detect vehicle from a moving camera. The detection component involves a cascade of modules. First, road is detected Based on edge detection and blob analysis, then our proposed motion detection (simplified Lucas-Kanade algorithm) is used to obtain motion vectors from two consecutive input frames. Third, frame differencing method Based on obtained motion vectors is applied to generate differenced image. At last, moving vehicles are detected Based on blob analysis.