A Method of Target Fusion Detection Based on the Multi-fractal Analysis

A target detection method is proposed which is based on the multi-fractal spectrum characteristics of the remote sensing image. The method uses multi-fractal spectrum to reflect the image global features and effectively suppress noise characteristics at the same time. The membership function is constructed by multi-fractal spectrum of visible and SAR images to indicate the uncertainty of their sensors information. The trade-off between the intersection operation and union operation of the feature level fusion detection is achieved by the membership function, in order to reduce the false alarm rate and the missed alarm rate. The experimental results show that the method can combine the respective advantages of SAR and visible images, and effectively improve the accuracy of detection.