A novel approach for automatic aircraft detection

The automatic recognition of planes in aerial images is an important application in the image analysis field. However, it remains a problem despite many years of work due to the arbitrary original poses and the variation in the shapes of planes. This paper proposes a novel approach for automatic aircraft detection based on statistical theory and common features of different kinds of planes. Experiments show its advantage in both speed and accuracy. It is another application of statistical theory applied to image processing.