In this paper , deconvolution and linearized inversion of receiver function are improved ; multi - channel deconvolution and wavelet inversion are developed for receiver function . the elastic wave motion equation is applied in the simulation and migration of receiver function in lateral inhomogeneous media , multi - grid algorithm is introduced in numerical modeling of elastic wave motion equation , and phase - delay boundary condition is also provided to absorb boundary reflection 在此基础上，重点研究和发展了横向非均匀介质中，基于波动方程的数值模拟与偏移成像方法，首次将多重网格算法引入到弹性波方程的数值模拟和接收函数的偏移成像，还发展了一种延迟边界方法，以消除人为边界反射的影响。
The clues and techniques to improve the resolution of data processing are 1 ) removing the affection of near surface factors on data acquisition , including static correction and noise eliminating technique etc , 2 ) broadening the frequency band of pre - stack seismic data , including amplitude compensation and pre - stack deconvolution etc , 3 ) improving s / n ratio of high frequency and weak reflection signals through multiple coverage with high precision velocity analysis , residual static correction and high - order nmo , raising " dead line of high frequency reflection " , and 4 ) identifying post stack effective reflection frequency band , enhancing dominant frequency and so on 主要包括：消除近地表因素对资料采集的影响，包括静校正和噪音压制技术等；叠前拓宽地震资料领带，包括振幅补偿和叠前反褶积技术等；借助高精度速度分析、剩余静校正和高次项nmo技术，通过多次覆盖叠加，消除资料中随机噪声干扰、提高资料信噪比，尤其是高频端反射信息的信噪比，改善高频反射和弱反射信号信噪比，提高“高频反射死亡线” ；叠后有限反射颇带识别、抬高优势频率等。
On the condition of the missing pixel information of the observed image during the process of imaging , based on the analysis of the degradation principle and establishment of the math model as well as discrete expression , this paper incorporates optimization method into the process of blur image restoration and proposes the deconvolution methods of non - zero background motion blurred image based on conjugate gradient optimization iteration 摘要针对成像过程中观测图像存在像素灰度信息丢失的情况，在分析此类图像的退化原理及建立其数学模型和离散化表达式的基础上，将最优化方法中的共扼梯度法引入图像复原过程中，提出基于共轭梯度优化迭代的非零背景运动模糊图像去卷积方法。
In this paper , based on the study of single channel deconvolution , we made a research in double channel deconvolution . there are three cases when we find its results by dfs algorithm . the first is that z transform expression of results have no polar on the identity circularity . the second is that denominator have zero on it while molecular have zero to counteract . the third is that z transform expression of results have polar on it , which is well known that deconvolution kernel has no spectral inverse 本文在对双通道信号卷积反演求解时采用较为简便的dfs算法，在求解过程中会遇到的三类情况，第一类是解的z变换表达式在单位圆上无极点，第二类是z变换表达式分母有零点但分子有相应的零点作抵消，第三类是z变换表达式在单位圆上有极点(即称之为卷积核无谱逆) 。
The problem discussed in this paper is to separate two ( or more ) input signals from observed signals which are generated by passing input signals through diffrent unknown multi - input multi - output linear systems . it is proved that the input signals can be separated when they are independent identitically distributed ( i . i . d ) signals . a new algorithm for multi - input multi - output blind deconvolution via maximum entropy is presented which needs no information about the input signals and mixing filters 本文研究的问题是从观察信号中分离出两个(或者更多个)输入信号，其中每一组信号分别通过不同的未知多输入多输出线性系统.本文证明了当输入信号是两两相互独立的独立同分布信号时可以分离出输入信号，并导出了基于最大熵的多输入多输出盲解卷新算法.这个算法不需要任何关于输入信号和混合滤波器的先验知识
Multi - channel deconvolution , based on multi - channel signal analysis , is presented to recover receiver function from three - component teleseismic p waveforms clustered within certain backazimuth and epicenter distance , to increase the resolution of receiver function . wavelet inversion is also introduced to broadband receiver function , to solve for nonuniqueness cased by linear approximation of nonlinear problem , and broadband receiver function is decomposed into multi - resolution versions , and the inversion result for the low - resolution version is iteratively taken as the initial model for the high - resolution , and gradually obtain the inversion model for broadband receiver function 为克服广义线性反演对初始模型的依赖性，并保持其快速收敛的特点，本文首次将小波变换引入到接收函数波形反演，将接收函数分解到不同的分辨尺度上，分别进行广义线性反演，并将低分辨率的反演结果作为高分辨率反演的初始模型，渐近地获取宽频带接收函数的模型，克服了非线性问题线性化带来的非唯一性。
In view of characteristics of seismic data of gobi - sand dune areas , this paper , through various data - processing means such as quality control , high - pass filltering , dip angle filltering , field static correction , surface consistency deconvolution , precise velocity analysis , denoising before stack , multiple iterative operation of residual static correction , denoising after stack and modification after shift , has raised the data quality of seismic profiles in low signal - to - noise ratio areas and managed to form the seismic data - processing technique for such complex areas as gobi , foreland and dune 摘要针对戈壁、沙丘地区地震资料的特点，通过对原始资料进行质量控制、高通滤波、倾角滤波、野外静校正、地表一致性反褶积、精细速度分析、叠前去噪、剩馀静校正多次迭代、叠后去噪和偏后修饰等处理，提高了低信噪比地区地震剖面的资料质量，摸索出了一套戈壁、山前、沙丘等复杂地区地震资料处理技术。
According to characteristics of the seismic data from the low snr region with complex surface , this paper is started from solving the static correction problem and reasonable eliminating all kinds of disturbance in the seismic data from the low snr region with complex surface . through the whole process of seismic data processing which includes a series of processing methods that are suitable for the low snr region , namely , from the choosing of the floating base - level , the static correction in the field and indoors , the eliminating of all kinds of noise before and after stacking , the velocity analysis with high - resolution , the reasonable techniques of deconvolution before stacking and wavelet processing after stacking , to the method choosing of the high - resolution stacking and the reasonable and accurate offset imaging , a set of the complete and effective flow for processing seismic data from the low snr region with complex surface are finally formed , which can meet the need of explo 本文针对复杂地表低信噪比地区地震资料的特点，以解决复杂地表低信噪比地区地震资料静校正问题及合理剔除各类干扰为出发点，在整个地震资料处理过程中，从浮动基准面的选取、野外及室内静校正、叠前叠后各类噪音的去除、高精度速度分析、合理的叠前反褶积及叠后子波处理技术、到选用高精度的叠加技术及合理准确的偏移成像方法等一系列适合于复杂地表低信噪比地区的处理方法，最终形成一套较完整且有效的针对复杂地表低信噪比地区地震资料的处理流程。
Thirdly , two sampling oscilloscope , agilent 86100a , are connected successfully . the configuration is introduced detailed and the simple procession to separate the kick - out pulse from the measured data is given . and the argumentation related to increase the precision , remove the noise and the deconvolution are conclude 第三、详细介绍了两台取样示波器agilent86100a实际对接的实验方法，并对实际测量数据进行了初步的处理，然后针对如何提高精度、去除噪声以及对反卷积的基本技术进行了有关的讨论。
But few involve in multi - channel deconvolution . in this disquisition , on the basis of one - channel deconvolution , we first investigate in the solution ' s number , quality , fast arithmetic ( esp dps arithmetic ) , error ' s processing , which belong to two - channel deconvolution . last , we obtain the method and conclusion to resolve the problem of two - channel deconvolution 本文是在单通道信号卷积反演问题研究结果的基础上，对双通道信号卷积反演的解的数目及性质、解的快速算法( dfs算法)及误差进行了研究，从而得出解决双通道信号卷积反演问题的方法和结论。
