嵌入式异构模型下低分辨率γ能谱实时解析方法研究

The extraction of spectral parameters is very difficult because of the limited enegry resolution for NaI(TI) and LaBr3(Ce) gamma-ray detectors, and it takes more time for system response to finish processing complicated algorithms. Then, An Embedded Heterogeneous System with multiple output nodes is proposed. By studying co-scheduling algorithms between hardware and software, communication mechanism among the nodes, high-speed data transmission between computing node and control node become available.For statistical fluctuation of radioactivity under complex environment,some methods are proposed to solve the problem. These methods include adopting some energy-spectra real-time decomposition algorithm such as combined peak search and overlapped peak decomposition based on deconvolution,peak net area calculating based on least square method,nuclide identifying and nuclide activity computing based on Optimal Linear Associative Memory neural network(OLAM), obtaining real-time various parameters of overlapping peaks and correcting counting loss which is produced from dead time and random increment effect.On this basis,key technical issues of portable spectrometer which has some key features such as real-time nuclear spectral data processing, control of spectral parameters and high-speed transmission of spectrum data will be solved. Finally a real-time parsing model for low-resolution γ spectrometry based on embedded heterogeneous system is construted. At last, in academic aspect, the new instrument will contribute to developing theory of intelligent nuclear instrument. In application aspect, the device will be widely used to mineral exploration and radioactive monitoring in Jiangxi province.

针对γ能谱测量中（NaI(TI)或LaBr3(Ce)）探测器有限的能量分辨率致使谱参量提取复杂化，进而导致谱处理算法对系统响应产生时延的难题。本项目拟构建多节点输出的嵌入式异构系统实验模型，研究软硬件协同任务调度算法与各个节点间的通信机制，满足运算节点与主控节点之间实时谱数据传输。综合考虑实际应用环境中存在放射性的统计涨落，将重点采用反卷积联合寻峰与重峰分解、最小二乘拟合求取净峰面积、最优线性联想神经网络（OLAM）核素识别及活度计算等实时能谱解析方法；提取实时获取重叠峰的各个参数，校正死时间与随机加和效应而产生的计数丢失。解决实时能谱数据处理、谱峰参数控制、谱数据高速传输等软硬件于一体的便携式谱仪关键技术问题。研制适用于低分辨率γ能谱实时解析的嵌入式异构实用模型。为应用在江西丰富的矿产资源勘探及放射性环境监测等领域中新型智能化核仪器的理论与实践研究奠定坚实基础。

在γ射线能谱测量中，NaI(Tl)闪烁探测器具有探测效率高、价格低廉等优势被广泛应用，但由于其探测器的能量分辨率有限，导致准确解析该类探测器测得的仪器谱成为一个关键性难题。本项目不断探索核能谱测量技术与嵌入式系统交叉融合，构建了异构系统结构实验模型，完成了基于ARM/DSP与ARM/FPGA异构嵌入系统架构下的多道脉冲幅度分析器研制，提出了基于NaI(TI)闪烁探测器谱仪获取的复杂仪器谱高分辨解析方法：（1）运用蒙卡方法模拟γ光子与NaI(Tl)探测器作用，建立了不同能量γ射线与探测器作用的蒙卡响应矩阵，采用改进的Boosted-Gold迭代算法求解响应矩阵参数，通过解析模拟多特征能量伽玛谱线和模拟合成重峰谱线等试验，最终实现在该响应矩阵下反演解析被测样品U、Th及U-Th混合源实测的γ解析谱。该解析方法够较好区分能量相近特征峰，改善由于重峰导致活度分析的误差水平，效果显著。（2）根据NaI(TI)闪烁探测器物理特性及成谱过程，不同能量γ光子在探测器中的响应对应光电峰的半高宽不同，且光电峰峰形近似高斯函数。通过分辨率刻度关系确定探测器的半高宽参数，并在放射源与γ谱之间构建针对该探测器的通用高斯响应矩阵，进而应用该响应矩阵反演解析该探测器获取的γ仪器谱，同时，采用改进的Boosted-Gold迭代算法求解病态矩阵，提取重叠峰的各个参数，最终实现在对应的响应矩阵下对每一次核事件产生参数时快速准确解析。利用高斯响应矩阵解析多重合成峰及U、Th、K实验谱的测试表明，解析的结果是测量谱线在该响应矩阵下对应的能量点或接近于理论上物理谱线的解，达到了放射性核素定性定量准确分析的目标。解决了能谱数据快速解析、谱峰参数控制、谱数据传输等便携式能谱仪关键技术问题，为矿产资源勘探及放射性环境监测等领域中新型智能化核仪器研究奠定坚实理论基础。. 本项目实施期间在国内外期刊上共发表学术论文13篇，其中期刊论文9篇，会议论文4篇，SCI检索3篇，EI/ISTP检索5篇，核心期刊4篇；授权发明专利4项，授权实用新型专利5项；获批软件著作权3项；培养了5名硕士研究生毕业。