探测因子优化的多尺度形态滤波器去噪方法

采用形态滤波器对信号进行去噪，结构元素尺寸的确定是最重要的一环。然而绝大多数的确定方法都是凭借经验或者同理想波形对比来优化结构元素尺寸，前者缺乏精确性，后者则缺少实际意义。针对这些问题，提出了一种通过探测因子函数来确定结构元素尺寸的新方法。首先，通过一个探测因子函数获得信号的尺寸信息。然后，对获得信息求导以找到合适的结构尺寸。最后采用多尺度金子塔形态滤波器来实现去噪。随机信号的仿真去噪以及激光脉冲雷达回波信号的去噪实验表明，该方法有着比较好的去噪效果，并且与其他常用形态滤波方法相比，既不依赖经验，也不引入理想波形。因此该方法有着非常好的前景。

De-noising method based on multiscale morphological filter optimized by detection factor

Using the morphological filters for de-noising, the determination of the structuring element's size is the most important step. However, most of the determination methods depend on either the experience or a comparison with the ideal waveform for optimizing structuring element's size, the former is lack of accuracy, the latter is lack of practical significance. Aiming at these problems, the paper proposes a new method that uses a detection factor function determinate the structuring element's size.Firstly, the message of signal structuring size was obtained by the use of detection factor function.Secondly, with the derivation of the message, proper sizes of the structuring elements were acquired.Finally, the multiscale pyramid type filters was used to de-noise the signal. Through the simulation of random signal de-noising and the de-noising experiment of lidar return signals, this method demonstrated better de-noising effect. Comparing with other usual morphological filtering methods, it neither depends on experience nor the introduction of the ideal waveform. Therefore, this method has a promising future.