K-SVD字典在航空伽马谱数据降噪中的应用研究

Research on K-SVD Dictionary in Airborne Gamma-ray Spectrum Data Denoising

为降低航空伽马谱数据中的噪声,消除统计涨落对数据产生的影响,开展了K-SVD训练字典的降噪应用研究。根据稳定性测试所采集到的航空伽马谱数据,采用K-SVD训练字典进行了降噪实验,并将信噪比、皮尔森相关系数以及均方差等值作为降噪效果的评价指标,与NASVD以及S-G算法进行了降噪效果对比分析。实验结果表明,经过K-SVD字典降噪后的上下测窗谱数据与真值数据之间的相关系数分别达到了0.9983和0.9999,其整体降噪效果要优于NASVD与S-G算法。进一步利用不同降噪方法对实际测区数据进行处理,结果表明K-SVD算法能够有效滤除噪声并还原地质体特征。分析认为K-SVD算法能够有效降低统计涨落对原始波形带来的影响,在提升波形平滑性的同时使得数据的波峰波谷等特征得到较好的还原,在航空伽马谱数据处理中具有较好的降噪效果。

In order to reduce the noise in the airborne gamma spectrum data and eliminate the influence of statistical fluctuations on the data,the K-SVD training dictionary is applied to the noise reduction of the airborne gamma-ray spectrum data.The airborne gamma-ray spectrum data in xxx area are used as experimental data,K-SVD、NASVD and S-G are used as processing algorithms.Then using signal-tonoise ratio,person correlation coefficient and mean square error as the evaluation index of noise reduction effect.The correlation coefficients of the upper and lower window spectrum data that denoised by K-SVD dictionary and the true value data are 0.9983 and 0.9999,respectively.Its noise reduction effect is better than NASVD and S-G.Then different noise reduction methods are used to process the data of the actual survey area data,and the results show that K-SVD algorithm can effectively filter out the noise and restore the characteristics of geological bodies.The experimental results prove that the K-SVD algorithm can effectively reduce the statistical fluctuations on the original waveform.And the smoothness of the waveform is improved at the same time the impact characteristics are preserved well.