利用波形去噪算法的森林最大冠顶高估测研究

Research on ICESAT-GLAS waveform denoising algorithm and forest maximum crown height estimation

针对大光斑激光雷达回波信号噪声影响森林冠顶高估测精度,且回波分析法判定回波位置受限于平坦地区的问题,利用高斯低通滤波和小波去噪两种方法对GLAS波形进行去噪处理,提出了结合均方根倍差法和回波分析法来判定回波位置的有效算法。经小波去噪后信号的信噪比23.360 704,均方根误差为0.000 233 3,经均方根倍差法和回波分析法相结合来判定回波位置估测的冠顶高结果与实测结果相关性系数r值为0.864,效果均优于高斯低通滤波去噪。基于GLAS回波数据实验结果表明:小波去噪较好地实现了对回波信号的去噪处理,均方根倍差法和回波分析法相结合,实现了对坡度相对较大地区的GLAS波形的回波开始位置和地面回波位置的准确判定,对森林冠顶高的精确估算具有重要意义。

Aiming at the problem that the large spot laser radar echo signal noise influence forest crown height estimation accuracy and the echo analysis method to determine the echo position is limited to the flat area, in this paper, the Gaussian low-pass filtering and wavelet denoising methods are used to de- noise the GLAS waveforms, and an effective algorithm for determining the echo position is proposed by combining the root mean square method and the echo analysis method. The signal-to-noise ratio（SNR）of the signal after wavelet denoising is 23. 360 704, the root mean square error is 0. 000 233 3, the correla- tion coefficient between the crown height and the measured result is 0. 864 after the echo position is deter- mined by the combination of the root mean square method and the echo analysis method, the effect is bet- ter than Gaussian low-pass filter denoising. Based on the results of GLAS echo data, it is shown that wavelet denoising is a good way to realize the denoising of echo signal, the method of root mean square method combined with echo analysis realize realizes the accurate determination of the echo starting position and the ground echo position of the GLAS waveform in the relatively large slope area, which is of great significance to the accurate estimation of the crown height.