NASVD方法在航空γ能谱测量中的成图效果讨论

Airborne gamma ray spectrum mapping effects with NASVD de-noising

NASVD方法(noise-adjusted singular value decomposition)是一种基于多元统计分析来降低测量固有噪声的数据处理方法,通过将多道航空γ能谱仪采集到的大量测点构成的谱线矩阵分解为若干特征向量,再舍弃仅反映噪声的部分特征向量,利用噪声较少的特征向量对谱线矩阵进行重组。重组后,谱线矩阵中提取出的钾、铀、钍窗的计数率噪声均显著降低。笔者主要通过一些实际测区数据的处理结果来讨论聚类NASVD方法在不同情况测区的实际降噪效果。使用该方法分别对不同比例尺及飞行高度的实际测量数据进行了成图处理,并对处理结果进行了比较。在小比例尺航空γ能谱测量中,本方法优势并不十分显著。但在大比例尺,低高度的测量中,NASVD方法在空间分辨上较传统简单滤波方法有着非常明显的优势。

NASVD（ Noise adjusted singular value decomposition） is a spectral component analysis procedure for the removal of noise from gamma ray spectra. The procedure transforms observed spectra into orthogonal spectral components. Noise is removed from the observed spectra by rejecting noise components（ high order orthogonal components） and reconstructing the spectra from lower order components which are correlated with the signal. The noise of all channels is significantly decreased by this procedure. Three Airborne datasets of survey are processed to discuss the mapping effect with NASVD de-noising. The advantage of large scale mapping with NASVD is much significant compared with traditional average filter. Nevertheless,in small scale situation,they almost show the same result.