Target Detection in SAR Image Based-on Wavelet Transform and Fractal Feature

This paper introduces a small target detection algorithm combined with wavelet transform and high-order fractal feature for SAR image. The new algorithm is composed of three steps :(1) decompose the original image using wavelet transform and (2) extract the wavelet coefficients in low frequency and calculate its high-order fractal parameter to judge the target region, and (3) inverse wavelet transform to obtain the result image. Results are presented for two fractal methods (high-order fractal feature and fitting error). Our analyses show that the effect of using high-order fractal feature is better than fitting error. Based on the experiment result available to us, the small target is detected exactly from the SAR image.