航空伽马能谱异常信息提取方法研究

发布时间：2018-03-24 09:26

本文选题：航空伽马能谱　切入点：傅里叶变换　出处：《成都理工大学》2016年博士论文

【摘要】：航空伽马能谱测量(Airborne Gamma-ray Spectrum Survey)是将航空伽马能谱仪安装在飞机或其他飞行器上,测量空中伽马射线能谱,从而确定地表介质中放射性元素钾、铀和钍含量。它不仅用于寻找放射性矿以及与放射性元素有关的铀矿、钾盐、稀有元素和多金属矿床。此外还用于岩性地质填图和辐射环境监测等诸多领域。因其效率高、成本低、不受地形等因素限制而得到广泛应用。航空伽马能谱测量数据的后处理是航空伽马能谱测量的常规性且非常重要的工作内容,其中异常信息提取是航空伽马能谱测量应用的核心与热点技术问题。常规的航空伽马能谱异常信息提取沿用的是地球化学异常信息提取方法,该方法基于传统统计学的异常识别,其假设前提是数据必须服从正态分布或者对数正态分布,而这种完美的数据在实践中是几乎不可能出现的。而为保证所套用模型的准确,就需要对原始数据进行剔除,并依次循环剔除直至无离群值。但剔除后原始数据已不完整,无法完整客观还原所测区域的放射性地质特征。此外,统计学方法只依赖平均值、标准差等统计学参数来评价航空伽马能谱异常信息,而没有考虑这些数据本身所具有的空间频域属性和所携带的地球物理场特征。基于以上几点,本文提出了几种以空间频率域滤波和地球物理场分解为主要手段的航空伽马能谱异常信息提取方法,开展了如下研究工作:一、运用常规统计学方法对测区的航空伽马能谱数据进行处理,识别和提取异常信息,并进行异常评价与分析;二、将信号处理中一维傅里叶变换方法引入航空伽马能谱数据的后处理,从频率域角度出发,研究航空伽马能谱数据功率谱密度特征和航空伽马能谱背景和异常的空间傅里叶滤波分离技术;三、探讨基于不同小波基函数的航空伽马能谱数据处理方法和小波逐级分解的数据压缩技术;构建小波包残差面重构算法识别测区内不同种类型金属矿点;四、利用多重分形理论研究航空伽马能谱数据的多重分形特征,并根据其特征构造基于频域多重分形的航空伽马能谱数据逐层滤波模型,建立相应滤波方法以消除条带状异常;五、设计旨在分离和识别航空伽马能谱弱异常的样本熵矩阵算法,确定适用于航空伽马能谱数据处理的最佳熵平均窗口宽度值。通过以上内容的研究工作,对航空伽马能谱的空间频率域特征进行解析,取得了如下研究成果:(1)在分析航空伽马能谱数据特征的基础上,运用频率统计、相关聚类分析和主成分分析等常规方法处理测区航空伽马能谱数据,得到铀异常面积7.11km2,占测区面积的22.97%,异常极差为20.76,衬度值为4.35,异常区域较好的反映测区内花岗岩、沉积岩和砂岩等岩性分布,且与测区内铅锌矿点伴生的放射性异常相对应,但无法消除测量误差引起的条带状异常分布,为本论文提出的航空伽马能谱数据处理与异常信息提取技术的应用效果提供了比对依据。(2)基于一维傅里叶变换的功率谱密度与频率分布特征,提出了航空伽马能谱数据频率域异常参数划分方法和空间傅里叶滤波技术:低于0.02Hz频段内功率累积量不随频率增大而变化,可认为是稳定地质体在频域上的响应,低通滤波提取的异常面积4.48km2,异常面积占比14.46%,异常极差25.91;0.02-0.15Hz频段功率谱累积量随频率增加而降低,为铀异常频率分布特征,带通滤波提取异常面积低至0.03km2,异常极差11.59,且位置与区内铅锌矿点吻合;空间傅里叶滤波技术能有效校正条带状异常,在0.05Hz频率下,异常面积占比下降至1.17%,异常极差为11.59,异常形态清晰。(3)创建了航空伽马能谱数据小波残差面的小波包重构算法,利用残差面小波重构准确识别出测区中铅锌矿和铁矿;提出了基于Haar、Daubechies(dbN)、Bior双正交等不同小波基和小波逐级分解的数据压缩与处理方法,有效地提高数据处理效率,得出测区的异常面积占比分别为7.27%、6.67%和9.47%,异常极差分别为99.03、100.92和105.73,具有实用价值。(4)基于多重分形理论和航空伽马能谱数据特征,提出并确定了空间域功率谱的分形间断点及其左右两侧的显著特征,得出间断点左侧功率谱不随分形频率增加而变化,间断点右侧斜率骤降至负值。利用这一特征,构造了基于频域多重分形的逐层滤波模型,建立了消除条带状假异常的中高频带通滤波方法,有效分离了测区的背景和异常。功率谱能量在低、中、高频不同尺度的异常面积占比为13.64%、6.41%和1.55%,且具有良好的异常形态。(5)设计并实现了分离航空伽马能谱异常信息的样本熵矩阵算法,构建了空伽马能谱铀背景场的零元素平面和非零点位异常,提出了适用于航空伽马能谱数据处理的最佳熵平均窗口宽度值n=30。圈定的异常区域与测区花岗岩分布一致,显著提高了弱异常的提取能力。
[Abstract]:Airborne gamma ray spectrometry (Airborne Gamma-ray Spectrum Survey) is the aviation gamma spectrometer installed on the aircraft or other vehicles, measuring aerial gamma ray spectrum, so as to determine the radioactive elements in surface medium potassium, uranium and thorium content. It is used not only for radioactive ore and uranium, and radioactive elements potassium salt rare element and polymetallic deposit. In addition to lithology geological mapping and environmental radiation monitoring and other fields. Because of its high efficiency, low cost, is not limited by the topography and other factors and are widely used. The airborne gamma energy spectral data postprocessing is the content of the work of the conventional airborne gamma ray spectrometry and very important among them, the anomaly information extraction is the core of the problem and the hot technology of airborne gamma energy spectrum measurement applications. The airborne gamma energy spectrum of the conventional anomaly information extraction using the geochemical anomaly The information extraction method, anomaly identification based on traditional statistical method, the assumption is that the data must obey normal distribution or lognormal distribution, and the perfect data is almost impossible in practice. And in order to ensure the accurate and apply the model, we need to remove the original data, and in turn cycle until no outliers. But the rejection after excluding raw data has been incomplete, unable to complete the objective to restore the measured area of radioactive geology. In addition, the statistical method only depends on the average value, standard deviation and other statistical parameters to evaluate the airborne gamma spectral anomaly information, without considering the spatial frequency properties of the data itself is and with the characteristics of geophysical field. Based on the above points, this paper presents the spatial frequency domain filtering and geophysical field of several decomposition of airborne gamma energy spectrum as the main means of Often the information extraction method, carried out research work as follows: first, the use of conventional statistical methods of airborne gamma survey area can be processed spectral data, identifying and extracting anomaly information, and analysis and evaluation of anomalies; two, the signal processing in the one-dimensional Fourier transform method in airborne gamma ray spectrum data postprocessing, starting in view of the frequency domain, the research of airborne gamma ray spectrum data and power spectral density characteristics of airborne gamma spectral background and anomaly spatial Fourier filtering technology; three, to investigate the different wavelet functions of airborne gamma energy spectrum data processing method and wavelet decompose the data compression technology based on wavelet packet construction; residual surface reconstruction algorithm recognition test in the same type of metal mines; four, using multi fractal theory to study the multifractal characteristics of airborne gamma ray spectrum data, and according to the characteristics of structure based on frequency domain The airborne gamma multifractal spectrum data layer can filter model, establish the corresponding filtering method to eliminate the banded anomalies; five, is designed to isolate and identify airborne gamma spectrum sample entropy matrix algorithm for weak anomaly, determine suitable for airborne gamma spectrum data processing are the best entropy flat window width value according to the above research work. The content of the analysis of the spatial frequency domain characteristics of airborne gamma ray spectrum, the research results obtained as follows: (1) based on the analysis of airborne gamma spectral data features, using frequency statistics, Correlation Clustering Analysis and principal component analysis method of conventional test area processing of airborne gamma ray spectrometric data, obtained uranium anomaly the area of 7.11km2, accounting for 22.97% of the survey area, the abnormal range is 20.76, the contrast value is 4.35, the abnormal area reflected in the survey area of granite, sedimentary rocks and sandstone lithology, and the survey area lead-zinc ore The associated radioactive anomalies, but cannot eliminate the measurement errors caused by the banded abnormal distribution, provides a comparison basis for application of airborne gamma is proposed in this paper can extract technology of data processing and anomaly information. (2) the power spectrum of the one-dimensional Fu Liye transform spectral density and frequency distribution is proposed based on Airborne gamma ray the spectral data in frequency domain and space division method of abnormal parameters of Fu Liye filtering: less than 0.02Hz band power accumulation does not change with increasing frequency, can be considered as the response of stable geological body in the frequency domain, low pass filtering 4.48km2 anomaly area extraction, anomaly area accounted for 14.46%, abnormal range 25.91 0.02-0.15Hz band power spectrum accumulation; with the increase of frequency and reduce the amount of uranium, abnormal frequency distribution, band-pass filtering and extract abnormal area of low range 11.59 to 0.03km2, abnormal, and the location and area of lead and zinc Mine anastomosis; spatial Fourier filtering technology can effectively correct the banded anomalies, abnormal area at the frequency of 0.05Hz, the proportion dropped to 1.17%, abnormal range was 11.59, abnormal morphology is clear. (3) created the airborne gamma ray spectrum data of the residual wavelet wavelet packet reconstruction algorithm using wavelet reconstruction to accurately identify residual surface measurement of lead and zinc the ore and iron ore; proposed based on Haar, Daubechies, Bior (dbN) biorthogonal different wavelet and wavelet decompose the data compression and processing method, effectively improve the efficiency of data processing, the measured area of the abnormal area accounted for 7.27%, 6.67% and 9.47%, respectively, and the abnormal range 99.03100.92 105.73, it has practical value. (4) multi fractal theory and airborne gamma spectral data based on the characteristics, proposed and identified a significant feature of fractal spatial domain power spectrum between the two sides of the breakpoint, the discontinuous The left power spectrum with fractal frequency increases, the slope discontinuity points dropped to negative. Using this characteristic, structure layer filtering model in frequency domain based on multifractal, established to eliminate false anomaly in high frequency band band-pass filtering method, effective separation of the surveyed area and background power spectrum anomaly. In the low energy, high frequency, abnormal area of different scales accounted for 13.64%, 6.41% and 1.55%, and with abnormal morphology good. (5) the design and implementation of sample entropy matrix separation algorithm of airborne gamma spectrum anomaly information, constructs the zero elements of plane air gamma energy spectrum and non uranium background field zero abnormal, was proposed for the airborne gamma ray spectrum data processing window of optimal entropy average width value anomaly area n=30. delineation and measuring distribution of granite area increased significantly consistent, ability to extract weak anomaly.

【学位授予单位】：成都理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：V248

【相似文献】