oa Comparison Between PCA and LFDA For BSR Detection Of Gas Hydrate Zones

Gas hydrate is one of the most important gas resources in the world, thus it can be a good fuel resource in the future. Therefore, its exploration is important for each country. Gas hydrates change the physical properties of the host sediments. Therefore, seismic attributes can be used for their identification. The purpose of this study is using pattern recognition tools to combine seismic attributes and produce some new ones which can separate hydrate points from nonhydrate ones. In this article two algorithms ‐ PCA and LFDA‐ are used and their efficiency in seismic pattern separation are surveyed. Results show LFDA can separate gas hydrates from nonhydrates better than PCA.