Title :
The pattern recognition of log facies using Walsh transform
Author :
Yang, Cheng ; Li, Xue-Yao
Author_Institution :
Dept. of Inf. Eng., Beijing Univ. of Posts & Telecommun., China
Abstract :
The recognition of log facies is very important in that of depositional facies. It is the key of recognition to find out the valid features that reflect the nature of the well-log. In accordance with the characteristic that the well-log is steady in a short term, a method of recognition of log facies is proposed which is based upon the applying of Walsh transform. In this method,. we pretreated the log curve by using the Walsh inversion to improve its vertical resolution whose aim is to obtain the real log curve to reflect the actual status in the well. Because the gist of marking the log facies is that the tendency of the well-log curve reflects the depositional variety of the lithology, a method is proposed that the short-term sequence spectrum of well-log curve is regarded as the eigenfunction. An experiment indicates that the short-term sequence spectrum of well-log curve is distinct from each other in different log facies. So, the proposed method is practical in projects.
Keywords :
geophysical prospecting; geophysical signal processing; geophysical techniques; pattern recognition; Walsh transform; eigenfunction; log facies; oil field; pattern recognition; petroleum prospecting; recognition; well-log; Computer science; Educational institutions; Eigenvalues and eigenfunctions; Fourier transforms; Image converters; Intrusion detection; Pattern recognition; Petroleum; Telecommunications; Well logging;
Conference_Titel :
Machine Learning and Cybernetics, 2002. Proceedings. 2002 International Conference on
Print_ISBN :
0-7803-7508-4
DOI :
10.1109/ICMLC.2002.1175409