Target fusion detection of remote sensing image based on the multifractal analysis

Multifractal characteristics show a strong correlation with surface textures of images. This paper describes the multifractal and how to calculate multifractal spectrum based on different measures, and has taken multifractal spectrum characteristics as the input of fuzzy C-Means clustering, finding that it can improve the accuracy of segmentation. The multifractal spectrum has reflect the global characteristics of images, in this case we have used the multifractal spectrums of images to complete the compromise fusion detection. The results show that the fusion detection based on multifracral has detected the target more integrated. Meanwhile the false detection dramaticly decreases.