改进辛普森-SNIP算法在航空γ仪器谱本底扣除中的应用

Application of improved Simpson-SNIP algorithm in background subtraction of airborne gamma-ray instrument spectrum

航空γ能谱测量中的本底计数主要是由于康普顿散射效应、γ射线在探测器灵敏体积内产生的小角度散射、电子噪声等干扰因素使能谱的能量区间的计数出现偏差。本底扣除是航空γ测量和解谱工作的重要任务之一。统计敏感的非线性迭代剥峰(Statistics-sensitive Nonlinear Iterative Peak-clipping,SNIP)算法具有简单而稳定的数学结构,自20世纪80年代起便被应用在γ仪器谱本底扣除中。但原始SNIP扣本底算法易受到峰宽参数的影响,且迭代过程中的收敛速度缓慢,导致运行时间延长。因此本文基于具有较高代数精度的辛普森公式的思想,对原始算法中的第二步SNIP迭代过程进行改进。结果表明:改进后的SNIP算法完整保存了原始算法简洁、高效的优点,同时选择了合适的变换宽度,准确度可提高41.8%,算法整体运行效率可提高2.53%,具有更优秀的本底扣除效果。

[Background] The background count in airborne gamma-ray spectrum measurement is mainly caused by Compton scattering effects, small angle scattering of gamma rays in the sensitive volume of the detector and interference caused by electronic noise in the energy range of the energy spectrum. Background subtraction is one of the important tasks of instrument spectral resolution in airborne gamma-ray measurement. [Purpose] This study aims to propose an improved background subtraction method to achieve accurate and efficient spectral analysis of airborne gamma-ray instrument spectrum. [Methods] As one of the best background subtraction algorithms, the statisticssensitive nonlinear iterative peak-clipping(SNIP) algorithm with simple mathematical structure and reliable background subtraction effect, was easily affected by the parameters of peak width, and lead to longer running time due to slow convergence speed of iterative process. Motivated by the idea of the Simpson formula with higher algebraic precision, the second step of the SNIP iteration in the original algorithm was improved by using three points of interval bisection for integral interpolation. Comparison of SNIP algorithm before and after the improvement was performed for background subtraction of airborne gamma-ray spectrum with different iteration times and appropriate width of transformation. [Results] The comparison results show that the accuracy and the overall operating efficiency of the improved algorithm are improved by 41.8% and 2.53%, respectively, compared with the original SNIP algorithm. [Conclusions] The improved Simpson-SNIP algorithm achieves fast convergence speed and better accuracy of calculation results while preserves the advantages of simplicity and efficiency of the original, hence better background deduction effect.