A target detection method for SAR image based on feature classification discrimination

This paper presents a target detection method for Synthetic Aperture Radar (SAR) image based on feature classification discrimination. Constant false alarm rate and extended fractal were used to detect targets in SAR images. Area and peak power ratio operators were used to discriminate targets and background clutter for eliminating a part of false alarms. Wavelet domain principal component analysis was applied to extract feature vectors from the images within detection window. Support vector machine was applied for classifying the extracted features vectors to discriminate targets and background clutter. Then, target detection was finished. ADTS data were used to verify and analyze the proposed method. The experimental results indicate after feature classification discrimination, the false alarms reduce significantly when the detection rate remains invariant. Therefore, the method presented in this paper is an effective method for SAR image target detection.