提高自然伽马测井曲线分辨率的正则化方法

应用正则化反演理论研究了提高自然伽马测井曲线纵向分辨率的正则反褶积处理技术，通过正则化因子对噪音和信号影响的分析和研究，给出了估计反褶积结果与地层真值偏差大小的新公式，提出了直接根据信噪比大小选择正则化因子的新方法．由于正则化因子保证使反褶积结果与地层真值偏差为最小，使计算结果的稳定性和精度均得到明显提高．此外，大部分测井记录的信噪比均可以很容易估计，从而大大提高了正则反褶积的适用性．最后通过理论模型实验和井场数据处理证明了正则反褶积技术是十分有效的．

REGULARIZED ALGORITHM FOR ENHANCEMENT OF VERTICAL RESOLUTION OF NATURAL GAMMA-RAY LOG

The regularized algorithm to enhance the vertical resolution of natural gamma-ray log is studied using the regular inverse theory. With analysis and research of the influence of regular factor on the noise and signal, the new formular to estimate the difference of decovoluted resultS and formation true value is giv en. The new approach to choose regular factor is advanced directely according to the ratio of signal to noise. Since such a regular factor is chosen by means of minimizing the difference of decovoluted results and formation trUe value, the stability and precision of the deconvolution are greatly improved. Besides, since the ratio of signal to noise can be easily estimated for majority of well logs, application of the regular deconvolution is greatly expanded. Finally the theory and aPProach are Proved effective and practical through the theoretical model experiments and field data processing.