Airport Monitoring System: Robust Airplane Extraction against Variable Environmental Conditions

We have developed an airport monitoring system which traces the movement of airplanes in the parking areas ofairports. For this system, we have developed an image processing method that can detect moving objects and determine the sizes of the objects under illumination changes. Conventional methoak adapted to changes in illumination are unable to correctly determine the size of objects because they can only detect the edges of the objects. We have therefore developed a two-stage background subtraction method which consists of local and global normalized subtraction. First, local normalized subtraction is applied to extract moving objects. Secondly, global normalized subtraction is applied to the area which encloses the detected objects to determine the size of the objects. With this method, airplanes can be detected in a stable manner amid changes in illumination. We have evaluated this method by using 140 hours of video images which contain changes in iNumination and confirmed 95% accuracy of airplane detection. This system is now in operation at the Kansai International Airport.