前后向平滑算法：算法的原理很简单，以五点三次平滑为例。取相邻的5个数据点，可以拟合出一条3次曲线来，然后用3次曲线上相应的位置的数据值作为滤波后结果。简单的说就是 Savitzky-Golay 滤波器 。只不过Savitzky-Golay 滤波器并不特殊考虑边界的几个数据点，而这个算法还特意把边上的几个点的数据拟合结果给推导了出来。-Before and after the smoothing algorithm: the principle of the algorithm is very simple, to 5.3 times smooth as an example. Take the adjacent five data points, you can fit out a curve 3 times, and then use the corresponding position on the third time the data value as the filtered results. Simply said is the Savitzky-Golay filter. However, the Savitzky-Golay filter does not specifically consider several data points of the boundary, and this algorithm also specifically to the edge of a few points of the data fitting results derived.