【摘要】 研究和分析了激光雷达目标回波与杂波背景的分形特性,以分数布朗运动(FBM)数学模型为基础,通过对不同信噪比(SNR)激光雷达回波的分析,初步证明了激光雷达回波具有布朗运动的特征。针对激光雷达杂波数据具有布朗运动的增量统计自相似性,激光雷达杂波能与分数布朗运动较好地匹配,因此可以采用布朗运动模型对激光雷达回波数据进行分析处理。当信噪比比较低时,杂波和含有目标的回波信号的分形维数比较接近,发生重叠,单靠单一的分形维数无法检测出目标。针对这一问题,分析杂波和目标信号在不同尺度上的分形维数,提出基于不同尺度分形维数的变化特征进行目标检测的算法,即基于多尺度分形维数的目标检测算法。理论分析和实验结果表明该方法具有较高的可靠性和准确性。更多还原

【Abstract】 The fractal characteristics of the laser radar′s echo and clutter are researched and analyzed.Based on the fraction Brownian motion(FBM) model and analysis of different signal-to-noise ratio(SNR) of the laser radar echo,it proves that the laser radar echo data can be processed has the characteristic of Brownian motion.Because the clutter of the laser radar has increment self-similar characteristic of Brownian motion,it can match the fractal Brownian motion.So the laser radar echo data can be processed by adopting the Brownian motion model.When the SNR of the target echo is small,the fractal dimension of the clutter and target echo is similar.The target can not be detected by the single fractal dimension.But the fractal dimension of the clutter and target is different in the different fractal scale.A method to detect the target by the characteristic of the different fractal dimension of the target and the clutter is advised.This method is called target detect algorithm based on multi fractal dimension.The theory and experiment result shows that this algorithm is reliable and accurate.更多还原