The Multi-fractal Characteristics Analysis Of Sea Clutter And Target Detection

The study of sea clutter by broad research institutions and scholars has a long history, inorder to study various characteristics of sea clutter, extensive and in-depth analysis is also carriedout. These features cover the statistical characteristics, the time-frequency distributioncharacteristics, the chaos characteristics and fractal characteristics. Through the analysis of thesefeatures, the researchers could set up a corresponding model or calculate the corresponding indexin order to realize target detection under the background of sea clutter.This thesis begins with the statistical characteristics analysis of sea clutter which is related totime and space correlation, the statistical correlation characteristics is found insufficient, as aresult, the fractal and multi-fractal properties are studied in the following part. Belonging to thecategory of fractal and multi-fractal theory, the long range correlation opens up a new direction forthe study of sea clutter correlation characteristics. Firstly, the de-trending fluctuation analysis(DFA) for single distance location of sea clutter is carried out, then the scale index is calculated bythe improved algorithm, and target detection is achieved by the difference of scale index. The DFAmethod for single distance location also has great limitation, in order to study different distancelocation of sea clutter, de-trending cross correlation analysis (DCCA) is creatively introduced.DCCA is used to study the cross correlation between different sea clutter cells, different targetcells, sea clutter and target cells. And several important conclusions are drawn by a large numberof experiments. In the end, in order to achieve further study for multi-fractal characteristics of seaclutter, sea clutter data is tested through the method of multi-fractal de-trending fluctuationanalysis (MF-DFA）. The experimental results have proved that the sea clutter signal hasmulti-fractal features, further experiments have found that the size of the singular strength indexbetween sea clutter and target signal has great difference. Target detection under the backgroundof sea clutter is realized by support vector machine (SVM) according to the difference. The methods of DFA, DCCA and MF-DFA which are used for the study of sea clutter areinnovative. Each study is supported by experiments of real sea clutter data, so this research hasgreat theoretical and practical value.