磁异常信号检测与源定位方法研究

发布时间：2018-06-14 21:25

本文选题：磁异常探测 + 标准正交基分解算法　；参考：《电子科技大学》2015年硕士论文

【摘要】：我们生活的环境中由于地球这个大磁体的存在而处处分布着磁场,所以地球上的任何物体均会受到地磁场的影响。磁性物体或被地磁场磁化的铁磁性物质,以及运动导体由于切割地磁场产生涡流磁场等,都会对地磁场产生扰动,称为磁异常现象;基于磁异现象对磁性目标进行检测和定位的方法就是磁异常探测-MAD(Magnetic Anomaly Detection)。磁异常探测技术是基于基本的物理现象,具有很好的稳定性和通用性,并且是一种被动式的探测方法,在军事、地质资源勘查、考古、地震预警、空间探测、医学诊断和生物磁学等方面,有着广泛的应用。本文主要讨论该技术在航空物探方面的应用,主要工作包括以下几方面:首先,基于磁异常信号相对环境磁噪声是十分微弱的情况,选取了能有效的提高磁异常信号信噪比的标准正交基函数(OBF)分解算法。该算法是通过求解磁异常信号在所选择的基函数空间上的能量函数来达到提高信噪比的目的,在分析中也验证了对特征时间的估计方法。进一步分析得到,由于三个基函数的自相关函数不是理想的?形函数,而互相关函数也不是恒为零,所以得到的能量函数往往不理想。于是提出了一种将三个基函数组合成一个新的基函数的改进方法,这样可以改善能量函数的形状使其更像一个尖脉冲。最后讨论了平台机动对能量函数的影响。其次,考虑到在高斯白噪条件下标准正交基函数分解算法对磁异常信号的处理效果是最理想的,而非高斯白噪声对应的检测效果不尽如意的情况。对在环境磁噪声中占主要成分的地磁噪声,基于其功率谱密度满足1/f?0 2?????的非高斯特性,选取了一种基于自回归模型(AR)的白化滤波方法。对如海洋背景磁场这种一般的非高斯噪声,提出了一种基于小波变换的白化方法,并通过仿真以瑞利分布和指数分布为例验证了其白化的有效性。然后,将能量函数作为目标检测时的检验统计量与门限值相比较,如果能量函数在连续的一段时间的内均大于自适应门限值,即认为有目标。为了达到恒虚警率(CFAR)检测的效果,基于能量函数的分布特性,采用两侧单元平均选大(GOCACFAR)的恒虚警概率检测算法来确定自适应门限值。最后,将传统的标准正交基函数分解算法的应用条件进行了拓宽,由之前的平台和目标一个固定一个运动拓宽到两个可以以任意速度和角度运动。并据此提出了一种具有一定实际应用价值的定位方法:型定位方法。
[Abstract]:We live in an environment in which there is a magnetic field everywhere because of the existence of a large magnet, so any object on the earth will be affected by the geomagnetic field. Magnetic objects or ferromagnetic materials magnetized by geomagnetic field, as well as eddy current magnetic field produced by moving conductor due to cutting geomagnetic field, will disturb the geomagnetic field, which is called magnetic anomaly phenomenon. Magnetic Anomaly detection is a method to detect and locate magnetic targets based on magnetic anomaly. Magnetic anomaly detection technology is based on basic physical phenomena, has good stability and versatility, and is a passive detection method in military, geological resources exploration, archaeology, earthquake warning, space exploration, Medical diagnosis and biomagnetism are widely used. This paper mainly discusses the application of this technique in aeronautical geophysical exploration. The main work includes the following aspects: firstly, based on the fact that the magnetic anomaly signal is very weak relative to the ambient magnetic noise, The standard orthogonal basis function (OBF) decomposition algorithm which can effectively improve the signal-to-noise ratio of magnetic anomaly signals is selected. This algorithm can improve the signal-to-noise ratio by solving the energy function of the magnetic anomaly signal in the selected basis function space. The method of estimating the characteristic time is also verified in the analysis. Further analysis shows that the autocorrelation function of the three basis functions is not ideal? The form function, and the cross-correlation function is not constant to zero, so the energy function is often not ideal. An improved method of combining three basis functions into a new basis function is proposed, which can improve the shape of the energy function and make it more like a sharp pulse. Finally, the influence of platform maneuver on energy function is discussed. Secondly, considering that the processing effect of the standard orthogonal basis function decomposition algorithm is the most ideal under the condition of Gao Si white noise, but the detection effect of the non-Gao Si white noise is not satisfactory. The geomagnetic noise, which is the main component of environmental magnetic noise, is satisfied with 1/f?0 _ 2 based on its power spectral density. A whitening filtering method based on autoregressive model is selected. For general non-Gao Si noise such as ocean background magnetic field, a whitening method based on wavelet transform is proposed, and its whitening efficiency is verified by taking Rayleigh distribution and exponential distribution as examples. Then, the energy function is compared with the threshold when the energy function is used as the test statistic for target detection. If the energy function is larger than the adaptive threshold value for a continuous period of time, then the target is considered as a target. In order to achieve the CFAR detection effect, based on the distribution characteristics of the energy function, the CFAR probability detection algorithm of GOCACFAR is adopted to determine the adaptive threshold value. Finally, the application conditions of the traditional standard orthogonal basis function decomposition algorithm are broadened, from the previous platform and the target to one fixed motion at a time to two moving at any speed and angle. Based on this, a localization method with practical application value is put forward.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM936

【相似文献】