采用内点法的基追踪算法测井数据恢复

实际测井数据由于数据规模大、维数高等特点,在采集、传输、存储中很有可能会发生数据的缺失或损坏。为了提高数据质量,采用压缩感知中的基追踪算法对测井中的缺失数据进行恢复。先使用过完备字典对原始数据进行稀疏分解,采用基追踪算法将L0范数优化问题转化为L1范数问题;然后采用线性规划算法——内点法求解,从而实现数据恢复。通过实验与对比分析,结果表明基于基追踪的测井数据恢复方法精度高;且优于基于正交匹配追踪的数据恢复方法。

Logging Data Recovery via Basis Pursuit Algorithm of Interior Point Method

Due to the actual logging data having the large scale of data, high latitude, it is likely to occur data loss or damage in the collection, transmission, storage. In order to recover missing values, a filling missing data method based on Basis Pursuit algorithm is proposed, which includes using a over-complete dictionary for the original signal to sparse decomposition, then filling in the missing data by Basis Pursuit (BP) algorithm. Basis Pursuit (BP) algorithm transforms the 0 norm optimization problem into 1 norm problem and adopts linear programming. The experimental results in logging data show a filling missing data method based on Basis Pursuit is remarkable, and is much superior to that of Orthogonal Matching Pursuit (OMP).