勘探地震学数据处理中的三种反褶积技术

在地震数据中，反褶积是用来提高分辨率的必要手段，但同时往往会降低资料的信噪比，当地震资料不满足最小相位和白噪声的约束条件时，常规的反褶积方法也将不再适用。从这两个问题出发，混合相位未知脉冲最小平方反褶积，多分辨率地震信号反褶积，神经网络子波反褶积等三种改进的方法分另运用多次迭代，二进小波变换和神经网络技术，对常规方法的不足予以改善。

Three Techniques of Deconvolution in Seismic Data Processing

Deconvolution is a necessary means of improving seismic resolution in data processing,but in the same time it can decrease the signal noise ratio.Conventional deconvolution does not work when seismic data don't satisfy the constraints of minimum phase and white noise . Least squares deconvolution of the mixed phase unknown pulse,multiresolution seismic signal decovolution and neural network wavelet deconvolution can improve conventional deconvolution with iterative method,dyadic wavelet transform and neural network technique.