The Sea Clutter Fractal Characteristic Analysis And Target Detection

Sea clutter refers to backscatter echo from local sea surface received by radar. It affected the radar performance in detection and positioning of targets on sea surface like iceberg, oil tankers, etc at a large extent. Therefore, it is critical to do research on characteristics of sea clutter for mitigating the effects of scattering and improving the target detection performance against sea clutter. Fractal is an important branch of nonlinear science, which has played a very good advantage in dealing with non-linear and non-stationary signal. Fractal has obtained the very good development and application in many fields, especially in the modeling of radar clutter and target detection. This paper focuses on the research of leading edge topics, that is, the application of fractal theory in the field of sea clutter and weak target detection. We adhere to the combination of theoretical research and validation on actual measured data. On one hand, we analysis the basic mechanism of sea clutter and the advantage of fractal theory in detecting the weak targets against sea clutter. On the other hand, IPIX real-life data are used to verify the effectiveness of the proposed method. This thesis concentrates on as follows:This thesis expounds fractal and its related theory, including the introduction of the fractal, self-similarity and self affine, the definition and solving method of common fractal dimension, and the basic concept of multi-fractal. Much attention is given to the applications of fractal and its related theory in the field of radar target detection. Single fractal and multi-fractal methods are utilized to do simulation and analysis on the fractal characteristics of the observed sea clutter data respectively.Extract two specific features of pure sea clutter from radar echo to compose a two-dimensional vector, then utilize convex hull training algorithm to obtain classification region by analyzing pure sea clutter. Extract two characteristic parameters as above from the unknown clutter by using the same method, discriminate whether there is a goal by judging whether the corresponding points are in the classification region or not. The method has been confinned with the data collected by Master IPIX radar at the east coast of Canada that the algorithm is superior to the contrast algorithms, that is, the detection effect is improved obviously under the same false alarm probability, which provides a new detection scheme for radar target detection.Further, analysis generalized Hurst index and multi-fractal spectrum of sea clutter and target echo on the basis of the measured data. It was found that the generalized Hurst index curve of the pure sea clutter and target echo separate a litter far when doing research on multiple sets of data, especially under the condition of low dimension, while the generalized Hurst index curves of the pure sea clutter in different range are relatively close. Based on these findings, we selected three multi-fractal characteristic parameters to quantify the differences between the two echo data. By using the SVM classifier to do classification experiment on target echo and pure sea clutter, the experimental result shows that both for the training sample and test sample, target detection based on multi-fractal characteristics of target unit and pure sea clutter unit correct detection rate is over90%. At the same time, we point out that the classification accuracy under different polarization modes exist a certain differences, the result shows that the multi-fractal characteristic parameters for testing can basically separated the target unit and pure sea clutter. As the experimental results showed, the method can achieve good effect and has a certain practical value.