核电站液态流出物排放，日本核污水排海和核潜艇事故等事件会造成大范围、长时间的海水放射性污染，所以需要一种应急监测方法实现对大范围海水γ核素的快速监测，及时响应放射性污染事件。常用的基于人工取样的实验室测量方法测量周期太长，而基于浮标的原位测量方法若要应用于大范围监测，其成本太高。利用船只搭载水下辐射探测器，采用走航测量的方法，是一种有效的解决方案。但在走航测量条件下，能谱测量时间很短，常用的探测器探测效率太低，无法及时响应放射性超标事件。且海水中γ核素含量极低，特征峰不明显，导致传统的核素识别方法的识别准确率下降。本文针对走航测量方法中存在的问题开展了水下辐射探测器阵列设计和基于RepVgg神经网络的核素识别算法研究。本文的主要研究内容和结论如下：

       （1）确定了水下辐射探测器阵列探测效率最高的结构参数，并研究了海水测量环境对探测效率的影响。通过蒙卡模拟构建探测器和海水源项模型，研究了探测器晶体规格、探测器间距和探测器晶体数量以及海水环境参数对探测效率的影响。结果表明，在NaI(Tl)晶体总体积固定为127.2立方英寸的情况下，选择3个Φ3"×6"或4个Φ3"×4.5"的晶体，组成探测器间距在30 cm以上的阵列，其探测效率最高。海水密度、温度和盐度对探测效率的影响可以忽略。在正常航行条件下，1个月需要清理一次探测器表面的海洋附着生物。

       （2）提出了一种基于RepVgg神经网络的核素识别算法，并研究了新算法的探测限。利用蒙卡模拟能谱与CeBr₃探测器获得的海水实测本底能谱相叠加，获取数据集作为神经网络的输入。从混合核素个数、能谱测量时间及核素活度浓度3个维度，研究了新方法的核素识别性能。测试结果表明，当混合核素个数小于10个，能谱测量时间大于300 s和核素活度浓度大于0.5 Bq/L时，新方法的识别准确率能够保持在90%以上，比CNN神经网络高1%~5%，比对称零面积法高20%以上。根据探测限的定义，模拟获得了核素识别方法的探测限，其随着核素个数的增加而增加，随着测量时间的增加而先减小后趋于平稳。

       （3）搭建了水下辐射探测器阵列实验装置，包括3个Φ3"×6" NaI(Tl)探测器和1个Φ2"×2" CeBr₃探测器，并开展了活度浓度测量实验和核素识别算法性能测试实验。实验结果表明，标准液体源中¹³⁷Cs和¹³³Ba的测量活度浓度与实际活度浓度偏差小于5%。走航测量实验中⁴⁰K的平均测量活度浓度为11.4 Bq/L与实验室分析结果偏差在3.6%到21.3%之间，验证了水下辐射探测器阵列的准确性。新算法对测量时间为20分钟的标准液体源能谱中的¹³⁷Cs的识别准确率为100%，比CNN高9.3%，比对称零面积法高41.7%。其对测量时间为5分钟的海水能谱中的⁴⁰K的识别准确率为100%，与其他算法性能相当，验证了新算法的可行性。

       本文设计了水下辐射探测器阵列，提高了探测效率，能够准确测量核素活度浓度；提出了基于RepVgg神经网络的核素识别算法，对短时间测量条件下的海水γ能谱具有较高的核素识别准确率，适用于走航测量。相关成果对大范围海洋放射性原位监测领域的研究具有意义。

The discharge of liquid effluents from nuclear power plants, the release of Japanese nuclear wastewater into the ocean, and nuclear submarine accidents could result in widespread and long-term radioactive contamination of seawater, necessitating a rapid monitoring method for gamma radionuclides over large oceanic areas. Traditional laboratory-based sampling methods have lengthy monitoring cycles, while buoy-based in-situ measurement methods are cost-prohibitive for large-scale monitoring. One potential solution is to equip underwater radiation detectors on ships and conduct survey measurements. However, under such survey conditions, the measurement time for energy spectra is short, and commonly used detectors have low detection efficiency, making it difficult to respond quickly to radioactive contamination events. Additionally, the low concentration of gamma radionuclides in seawater leads to a decrease in the accuracy of traditional radionuclide identification methods. This study addresses the challenges associated with survey measurements by designing an underwater radiation detector array and developing a radionuclide identification algorithm based on the RepVgg neural network, followed by experimental validation. The main research contents and conclusions are as follows:

(1) Impact of array structural parameters and marine environment on detection efficiency. The study used Monte Carlo simulations to construct models of the detector and source terms to analyze the effects of crystal specifications, inter-detector spacing, and the number of crystals on the detection efficiency of the array. The results indicated that when the total volume of NaI(Tl) crystals is fixed at 127.2 cubic inches, an array of three 3"×6" or four 3"×4.5" crystals, with inter-detector spacing greater than 30 cm, yields the highest detection efficiency. The addition of a 2"×2" CeBr₃ detector at the geometric center of the array has a negligible effect on detection efficiency. Marine environmental factors such as water density, temperature, and salinity have minimal impact on the detection efficiency. Under normal operational conditions, the detectors need to be cleaned of marine biofouling once a month.

(2) A novel algorithm based on the RepVgg neural network was proposed, and the detection limits of the algorithm were studied. The input data for the neural network was obtained by combining Monte Carlo simulated spectra and real seawater background spectra. The algorithm’s performance was evaluated based on three factors: the number of mixed radionuclides, measurement time, and radionuclide activity concentration. The results show that when the number of mixed radionuclides is less than 10, the measurement time exceeds 300 s, and the radionuclide activity concentration exceeds 0.5 Bq/L, the recognition accuracy of the new method is greater than 90%. This is 1% to 5% higher than CNN-based networks and more than 20% higher than the symmetric zero-crossing method. According to the detection limit definition, the detection limit of the algorithm increases with the number of radionuclides but initially decreases with increasing measurement time before stabilizing.

(3) An underwater radiation detector array, consisting of three 3"×6" NaI(Tl) detectors and one 2"×2" CeBr₃ detector, was constructed. Standard liquid source experiments, survey measurements, and sea well tests were conducted. The results demonstrated that the activity concentrations of ¹³⁷Cs and 133Ba in the standard liquid source deviated from actual concentrations by less than 5%. In the survey measurements, the average activity concentration of 40K was 11.4 Bq/L, with deviations ranging from 3.6% to 21.3% compared to laboratory analysis results. This validated the accuracy of the underwater radiation detector array. The RepVgg-based radionuclide identification algorithm achieved a 100% identification accuracy for ¹³⁷Cs in 20-minute spectra, outperforming CNN by 9.3% and the symmetric zero-crossing method by 41.7%. The algorithm also achieved 100% accuracy for 40K in 5-minute spectra, comparable to other methods. This confirms the feasibility of the proposed radionuclide identification algorithm.

This study designed an underwater radiation detector array that improves detection efficiency and energy resolution, enabling accurate measurement of radionuclide activity concentrations. The RepVgg-based radionuclide identification algorithm provides high identification accuracy for gamma spectra of seawater under short measurement times, making it suitable for survey measurements. The findings have significant implications for large-scale in-situ monitoring of marine radioactivity.

[1] 《2021 年中国海洋经济统计公报》发布[N]. 浙江国土资源, 2022(04): 18-19.

[2] 蔡福龙. 海洋放射生态学[M]. 北京: 原子能出版社, 1998: 5-8.

[3] 环境保护部(国家核安全局)有关负责人就《核安全与放射性污染防治“十三五”规划及 2025 年远景目标》答记者问[N]. 中国应急管理, 2017(03): 24-27.

[4] 生态环境部印发《“十四五”生态环境监测规划》[J]. 绿叶, 2022(Z1): 8.

[5] 蔡双雨, 宋术伟, 李馨楠, 等. 秦山核电站海域有害盐在带温核级材料表面沉积实验设计[J]. 实验技术与管理, 2024, 41(10): 28-34.

[6] 福建福清核电站工程进展情况[J]. 水泵技术, 2015(01): 52.

[7] 郭英来, 吴春元, 王虎, 等. 田湾核电站 2004-2022 年环境 γ 辐射剂量率连续监测分析[J]. 核安全, 2023, 22(06): 18-22.

[8] 张利民, 倪卫冲, 王彩霞, 等. 山东海阳核电站环境辐射本底航空测量调查[J]. 铀矿地质, 2016, 32(01): 36-42.

[9] 石同瑶,黄庆桥. 中国核电站反应堆技术路线的早期探索及现实启示——以秦山核电站为中心[J]. 中国科技论坛, 2024(12): 108-116.

[10] 李玉鑫, 张蔚华, 王忠杰, 等. 中国东北某核电基地流出物排放致公众剂量评价(2013—2020年)[J]. 辐射防护, 2023, 43(02): 155-165.

[11] 张博雅, 核电站常规液态流出物中核素在近海的迁移研究[D]. 2021: 45-50.

[12] 李宗明. 让切尔诺贝利核事故的警钟长鸣——纪念切尔诺贝利核事故 25 周年[J]. 核安全, 2011(03): 1-8.

[13] 陆遥, 技术解释视角下的三里岛核事故研究[D]. 2021: 10-20.

[14] 胡鹏磊, 陈惠芳, 袁龙, 等. 三大核事故应急撤离的问题与经验教训分析[J]. 中国辐射卫生, 2024, 33(06) : 681-685.

[15] 王储, 牛健植, 伦小秀, 等. 核电厂废水处理技术及水土保持措施的综述兼对福岛事故的建议[J]. 中国水土保持科学 2024, 22(6):10-28.

[16] 梁倩如, 王斌, 许浒. 核废水处理的现状与探索[J]. 当代化工研究, 2024(12): 80-82.

[17] 信萍萍, 陆燕, 张雪, 等. 福岛核事故放射性废水处理情况.中国核学会 2015 年学术年会[C]. 2015: 231-237.

[18] 潘寅茹, 福岛核废水 24 日起排海海洋渔业最先受创[N]. 第一财经日报. 2023(8): 1.

[19] 王兴春,伍浩松. 印度第二艘弹道导弹核潜艇服役[J]. 国外核新闻, 2024(09): 8.

[20] 苏玉顺, 陈彦, 刘承军, 等. 小型核应急医学救援力量建设及其在核潜艇核事故救援中的应用探讨[J]. 海军医学杂志, 2024, 45(04): 395-397.

[21] Bokor, I., S. Sdraulig, P. Jenkinson, et al. Development and validation of an automated unit for the extraction of radiocaesium from seawater[J]. Journal of Environmental Radioactivity, 2016, 151: 530-536.

[22] Gaur, S. Determination of Cs-137 in environmental water by ion-exchange chromatography[J]. Journal of Chromatography A, 1996, 733(1): 57-71.

[23] Povinec, P.P., J.J. La Rosa, S.H. Lee, et al. Recent developments in radiometric and mass spectrometry methods for marine radioactivity measurements[J]. Journal of Radioanalytical and Nuclear Chemistry, 2001, 248(3): 713-718.

[24] Huang, C.Y., J.D. Lee, C.L. Tseng, et al. A rapid method for the determination of 137Cs in environmental water samples[J]. Analytica Chimica Acta, 1994, 294(2): 221-226.

[25] Byun, J.-I., S.-W. Choi, M.-H. Song, et al. A large buoy-based radioactivity monitoring system for gamma-ray emitters in surface seawater[J]. Applied Radiation and Isotopes, 2020, 162: 109172.

[26] Tsabaris, C., C. Bagatelas, T. Dakladas, et al. An autonomous in situ detection system for radioactivity measurements in the marine environment[J]. Applied Radiation and Isotopes, 2008, 66(10): 1419-1426.

[27] Tsabaris, C.,A. Prospathopoulos. Automated quantitative analysis of in-situ NaI measured spectra in the marine environment using a wavelet-based smoothing technique[J]. Applied Radiation and Isotopes, 2011, 69(10): 1546-1553.

[28] 苏耿华, 海水就地γ能谱测量溴化镧探测器的技术研究[D]. 2010: 5-16.

[29] Song, J., P. Gong, P. Wang, et al. Unmanned stationary online monitoring system based on buoy for marine gamma radioactivity[J]. Applied Radiation and Isotopes, 2023, 191: 110528.

[30] 蒋帅. 冰海毒瘤——前苏联“共青团员”号核潜艇沉没始末[J]. 海洋世界, 2010(04): 52-54.

[31] 谷穗和. 火灾葬送“共青团”号核潜艇[J]. 水上消防, 2008(06): 46-47.

[32] Chen, C., J. Li, J. Wang. Marine radioactivity emission and monitoring after Fukushima Daiichi Accident[J]. Science & Technology Review, 2022, 40(17): 105-112.

[33] Long, S., E. Hayden, V. Smith, et al. An overview of the Irish marine monitoring programme[J]. Radiation Protection Dosimetry, 1998, 75(1-4): 33-38.

[34] Mertzimekis, T.J., P. Nomikou, E. Petra, et al. Radioactivity Monitoring in Ocean Ecosystems (RAMONES)[C]. Proceedings of the Conference on Information Technology for Social Good. 2021: 216-220.

[35] Tsabaris, C.,D. Ballas. On line gamma-ray spectrometry at open sea[J]. Applied Radiation and Isotopes, 2005, 62(1): 83-89.

[36] Androulakaki, E.G., C. Tsabaris, G. Eleftheriou, et al. Efficiency calibration for in situ γ-ray measurements on the seabed using Monte Carlo simulations: Application in two different marine environments[J]. Journal of Environmental Radioactivity, 2016, 164: 47-59.

[37] Chernyaev, A., I. Gaponov, A. Kazennov. Direct methods for radionuclides measurement in water environment[J]. Journal of Environmental Radioactivity, 2004, 72(1-2): 187-194.

[38] Baranov, I., I. Kharitonov, A. Laykin, et al. Devices and methods used for radiation monitoring of sea water during salvage and transportation of the Kursk nuclear submarine to dock[J]. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2003, 505(1): 439-443.

[39] Wedekind, C., G. Schilling, M. Grüttmüller, et al. Gamma-radiation monitoring network at sea[J]. Applied Radiation and Isotopes, 1999, 50(4): 733-741.

[40] Thornton, B., S. Ohnishi, T. Ura, et al. Continuous measurement of radionuclide distribution off Fukushima using a towed sea-bed gamma ray spectrometer[J]. Deep Sea Research Part I: Oceanographic Research Papers, 2013, 79: 10-19.

[41] Jones, D.G. Development and application of marine gamma-ray measurements: a review[J]. Journal of Environmental Radioactivity, 2001, 53(3): 313-333.

[42] 杨本, 宋. 浸没水中γ自动监测仪[J]. 现代科学仪器, 1999(1): 105-106.

[43] 侯胜利, 刘海生, 王南萍. 海洋拖曳式γ能谱仪在渤海的应用[J]. 地球科学(中国地质大学学报), 2007(04): 528-532.

[44] Zhang, Y., C. Li, D. Liu, et al. Monte Carlo simulation of a NaI(Tl) detector for in situ radioactivity measurements in the marine environment[J]. Applied Radiation and Isotopes, 2015, 98: 44-48.

[45] 于奇, 房宗良, 文其林, 等. 海水NaI-LaBr3双探测器在线监测γ能谱的MC模拟[J]. 核电子学与探测技术, 2018, 38(05): 639-644.

[46] 岳昌啓,牛德青. 放射性核素能谱分析方法综述[J]. 兵工自动化, 2023, 42(06): 44-47.

[47] 武雷超, 李江坤, 张光雅, 等. 基于自适应导数-高斯寻峰算法的自动稳谱研究[J]. 核技术, 2024, 47(08): 53-60.

[48] 李尚柏. 适用于X射线谱分析的一种寻峰方法[J]. 核电子学与探测技术, 1991(05): 296-299.

[49] 徐宏坤, 嫦娥一号月球伽玛能谱数据分析与处理方法研究[D]. 2011: 20-55.

[50] 王海, 黄宁, 何泽, 等. 改进对称零面积变换寻峰算法在拉曼光谱中的应用[J]. 光学学报, 2024, 44(03): 296-307.

[51] Stinnett, J., C.J. Sullivan, H. Xiong. Uncertainty Analysis of Wavelet-Based Feature Extraction for Isotope Identification on NaI Gamma-Ray Spectra[J]. IEEE Transactions on Nuclear Science, 2017, 64(7): 1670-1676.

[52] 王一鸣,魏义祥. 基于模糊逻辑的γ能谱核素识别[J]. 清华大学学报(自然科学版), 2012, 52(12): 1736-1740.

[53] 王崇杰, 张. 基于模糊识别的γ能谱定性分析[J]. 光谱学与光谱分析, 2003(05): 1028-1030.

[54] Jiangmei, Z., J. Haibo, Z. Qingping, et al. Pulse Signal Recovery Method Based on Sparse Representation[J]. Journal of Beijing Institute of Technology, 2018, 27(2): 161-168.

[55] 任俊松, 张江梅, 王坤朋. 基于SVD和SVM的核素识别算法[J]. 兵工自动化, 2017, 36(05): 50-53.

[56] 王博, 石睿, 刘敏俊, 等. 基于FPGA的卷积神经网络核素识别硬件加速方法研究[J]. 核电子学与探测技术, 2024, 44(02): 334-343.

[57] 周思益, 张江梅, 刘灏霖, 等. 基于GASF与MSVM的放射性核素识别方法[J]. 西南科技大学学报, 2023, 38(02): 78-84.

[58] 杜晓闯, 涂红兵, 黎岢, 等. 基于径向基神经网络仿真γ能谱模板库的核素识别方法[J]. 清华大学学报(自然科学版), 2021, 61(11): 1308-1315.

[59] 王瑶, 基于循环神经网络与注意力机制的核素识别研究[D]. 2021: 45-53.

[60] 何建平, 基于神经网络的快速核素识别与低本底解谱算法研究[D]. 2018: 5-30.

[61] Chen, L.,Y.-X. Wei. Nuclide identification algorithm based on K–L transform and neural networks[J]. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2009, 598(2): 450-453.

[62] He, J., X. Tang, P. Gong, et al. Rapid radionuclide identification algorithm based on the discrete cosine transform and BP neural network[J]. Annals of Nuclear Energy, 2018, 112: 1-8.

[63] He, J.-P., X.-B. Tang, P. Gong, et al. Spectrometry analysis based on approximation coefficients and deep belief networks[J]. Nuclear Science and Techniques, 2018, 29(5): 69.

[64] 刘向前, 基于卷积神经网络的海水放射性核素解析算法研究[D]. 2024: 25-50.

[65] 王浙琦, 基于蒙特卡洛模拟的海水γ能谱原位测量探测器关键参数的研究[D]. 2021: 60-70.

[66] 姚尧, 海水中放射性探测的蒙特卡洛模拟[D]. 2016: 5-33.

[67] 汪栋, 海水就地γ能谱测量实验与模拟研究[D]. 2015: 21-45.

[68] 吴祥余, 不同尺寸γ射线探测器响应函数及探测效率的蒙特卡罗模拟[D]. 2009: 55-60.

[69] 吴祥余, 朱迪, 葛良全, 等. Monte Carlo方法对不同尺寸NaI(Tl)晶体探测效率的刻度[J]. 核电子学与探测技术, 2009, 29(01): 207-210.

[70] 汪栋, 方方, 余松科, 等. 海水就地γ能谱的MC模拟[J]. 核电子学与探测技术, 2014, 34(09): 1045-1050.

[71] Ahmadi, S., S. Ashrafi, F. Yazdansetad. A method to calculate the gamma ray detection efficiency of a cylindrical NaI (Tl) crystal[J]. Journal of Instrumentation, 2018, 13(05): 5019.

[72] 袁之伦, 潘自强, 张艳霞, 等. 核电厂流出物低水平放射性核素监测问题的探讨[J]. 辐射防护, 2015, 35(01): 1-8.

[73] 张艳霞, 李锦, 柳加成, 等. 核电厂流出物放射性核素监测项目探讨[J]. 辐射防护, 2014, 34(06): 390-394.

[74] 海水比重与盐度互查表[J]. 福建水产, 2003(02): 38.

[75] 王保栋, 陈., 刘峰. 海洋中Redfield比值的研究[J]. 海洋科学进展, 2003(02): 232-235.

[76] Liang, M., T. Zhou, F. Zhang, et al. Research on convolutional neural network and its application on medical image[J]. Sheng wu yi xue gong cheng xue za zhi Journal of biomedical engineering = Shengwu yixue gongchengxue zazhi, 2018, 35(6): 977-985.

[77] Bhatt, D., C. Patel, H. Talsania, et al. CNN Variants for Computer Vision: History, Architecture, Application, Challenges and Future Scope[J]. Electronics, 2021, 10(20): 28.

[78] 熊章友, 李卫军, 朱晓娟, 等. 基于深度学习的短时交通流预测研究综述[J]. 计算机工程与应用, 2024, 24(02):50-64.

[79] 尹爱军, 吕明阳, 杨敏英, 等. RepVGG与CapsNet融合的轴承故障诊断[J]. 振动与冲击, 2024, 43(14): 301-307.

[80] Cai Z, Qiao X, Zhang J, et al. Repvgg-simam: An efficient bad image classification method based on RepVGG with simple parameter-free attention module[J]. Applied Sciences, 2023, 13(21): 11925.

[81] 万勇, 李继武, 李维汉, 等. 改进RepVGG的鸟类识别分类算法[J]. 中国新通信, 2024, 26(08): 22-24.

[82] 施辉记, 基于FPGA的RepVGG网络结构的设计与实现[D]. 2024: 55-60.

[83] 邓楷文,葛晨阳. 改进YOLOv5的轻量化红外交通目标检测[J]. 计算机工程与应用, 2023, 59(12): 184-192.

[84] Shen, X., T. Liu, Y. Wang, et al. Detection algorithm for parking space status based on convolution network structural re-parameterization[J]. Journal of Jilin University. Engineering and Technology Edition, 2022, 52(12): 2898-2905.

[85] 杨奎, 黄岩, 陈浩峰, 等. NaI(Tl)晶体对核废水中典型核素的γ能谱响应的MC模拟[J]. 中国高新科技, 2024(17): 40-41+47.

[86] 唐生达, 兰长林, 聂阳波, 等. Cs2～6LiYCl6:Ce探测器γ响应函数的高斯展宽及解谱研究[J]. 现代应用物理, 2021, 12(04): 22-28.

[87] 周银行,马玉刚. MCNP能峰展宽的NaI探测效率研究[J]. 核电子学与探测技术, 2007(06): 1061-1063.

[88] 张庆国, 尤景汉, 贺健. 谱线展宽的物理机制及其半高宽[J]. 河南科技大学学报：自然科学版, 2008, 29(1): 4.

[89] 俞怡, 海水中关键人工放射性核素快速检测技术研究[D]. 2018: 10-23.

[90] 黄德坤, 俞怡, 杜金洲, 等. 基于多种富集材料的海水中人工放射性核素快速监测[J]. 海洋技术学报, 2022, 41(04): 43-52.

[91] 倪甲林, 于涛, 黄德坤, 等. 核事故下海洋放射性应急监测方案探讨[J]. 应用海洋学学报, 2022, 41(02): 268-274.

[92] 李文红, 拓飞, 周强, 等. 核辐射应急情况下放射性核素的γ能谱快速分析方法[J]. 中国辐射卫生, 2019, 28(06): 684-687.

[93] 核动力厂核事故环境应急监测技术规范（HJ1128—2020）[S]. 北京: 中华人民共和国生态环境部, 2020.

[94] 石岩, 张颖颖, 吴丙伟, 等. 海水放射性传感器温度漂移校正方法研究[J]. 辐射防护, 2023, 43(03): 225-234.

[95] 庞巨手, 郑., 侯晓凤. 对称零面积变换法找峰[J]. 原子能科学技术, 1987(03): 270-279.

[96] 毕云峰, 李颖, 杜增丰, 等. 对称零面积变换结合L-M拟合自动识别重叠光谱峰[J]. 光谱学与光谱分析, 2015, 35(08): 2339-2342.

[97] 陈川, 葛良全, 谷懿, 等. 采用对称零面积法的高统计涨落谱线峰位解析研究[J]. 核电子学与探测技术, 2016, 36(02): 229-231.

[98] 钟琳. 基于小波变换的高精度多脉冲激光测距技术研究[J]. 应用激光, 2020, 40(01): 129-133.

[99] 李致远, 朱求志, 吴永焜, 等. 基于小波分析的无线传感网实时异常检测算法[J]. 南京师大学报, 2014, 37(01): 87-92.

[100] 高妍, 新型闪烁体探测器在地空核辐射测量中探测限的研究与应用[D]. 2020: 50-66.

[101] 沙连茂, 卫为强, 宣义仁. 环境放射性监测中探测限附近测量数据处理方法的探讨[C]. 全国放射性流出物和环境监测与评价研讨会, 2003: 216-225.

[102] 黄乃明. 低水平放射性测量中的探测限及其计算[J]. 辐射防护通讯, 2004(02): 25-32.

[103] 尹亮亮, 张耀, 孔祥银, 等. 液体闪烁计数法分析水中226Ra活度浓度及其不确定度评价[J]. 辐射研究与辐射工艺学报, 2023, 41(2): 98-104.

[104] 汪传高, 骆志平, 庞洪超, 等. 低水平放射性测量的判断限和探测限[J]. 中国辐射卫生, 2018, 27(06): 590-594.

[105] 刘晗晗, 姜孔华, 王志兵, 等. 田湾核电站附近海域海水中137Cs活度浓度分析与评价[J]. 辐射防护, 2017, 37(05): 369-373.