Automatic target detection in CCD image using adaboost

Abstract

In this thesis, a new fast detection and clutter rejection method is proposed for Automatic Target Detection System in CCD image. Fast computation is a critical point for defense application, thus we concentrated on the ability to detect various targets with simple computation. For fast detection, the proposed method uses a cascade structure of the Adaboost algorithm. The Adaboost algorithm was successfully used for face detection. The proposed method slightly modified the Adaboost method to detect tank targets when the training data set is not enough. A majority filtering is also proposed to reject clutters detected alone, which improves the detection rate. Experiments were performed with real tank images. The experimental results show that the proposed method is superior to the previous method and it is fast enough to be used in actual defense system.