No abstract

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.

No abstract

The noise of laser images is complex, which includes additive noise and multiplicative noise. Considering the features of laser images, the basic processing capacity and defects of the common algorithm, this paper introduces the fractal theory into the research of laser image denoising. The research of laser image denoising is implemented mainly through the analysis of the singularity exponent of each pixel in fractal space and the feature of multi-fractal spectrum. According to the quantitative and qualitative evaluation of the processed image, the laser image processing technique based on fractal theory not only effectively removes the complicated noise of the laser images obtained by range-gated laser active imaging system, but can also maintains the detail information when implementing the image denoising processing. For different laser images, multi-fractal denoising technique can increase SNR of the laser image at least 1~2dB compared with other denoising techniques, which basically meet the needs of the laser image denoising technique.multi-fractal; laser image; image denoising 1 Introduction

No abstract