研究目的

The problem of reliable detection of hydrocarbon deposits can be efficiently solved by simultaneously using several complementary search methods, including remote laser sensing of the Earth surface layer, which is currently one the promising methods, and passive seismic exploration. One of the problems of the use of the method of remote geochemical analysis of hydrocarbon deposits is the necessity of processing large data sets obtained by measuring Raman spectra using lidar systems.

研究成果

The results of the study showed that the use of computing facilities with parallel data processing at small volumes of the learning sample is inexpedient, since the main time is taken by data recording to the global processor memory. However, when processing relatively large data sets, a significant increase in both the neural network learning and spectrum analysis rates is observed.

研究不足

The use of computing facilities with parallel data processing at small volumes of the learning sample is inexpedient, since the main time is taken by data recording to the global processor memory.