Rock feature prediction model training method and device and rock feature prediction method

Abstract

The invention relates to the field of oil-gas exploration and development, in particular to a rock feature prediction model training method and device and a rock feature prediction method. Comprises: acquiring a historical rock sample; acquiring a rock component image of the historical rock sample and a corresponding rock component label; obtaining an atomic force microscope image of the historical rock sample and a corresponding mechanical information label; the rock component image, the rock component label, the atomic force microscope image and the mechanical information label serve as training samples, a rock feature prediction model is trained, the rock component label and the mechanical information label are different labels at the same position of the same rock, and the rock component label and the mechanical information label are different labels at the same position of the same rock. The rock component label and the mechanical information label have a corresponding relation. According to the scheme, the features in the rock image are rapidly identified, rapid positioning, accurate identification and mechanical property prediction of the rock components in the rock image are realized, the identification precision of the rock components is improved, and the test efficiency is improved.

Description

technical field [0001] This paper relates to the field of oil and gas exploration, especially a rock feature prediction model training method and device, and a rock feature prediction method. Background technique [0002] Mineralogical tests can identify the composition of rock minerals and predict their mechanical properties and other important information, and play an important role in petrological research. [0003] The traditional mineralogical test method is to polish the measured rock material with argon ions, obtain rock composition imaging and topographic information with an electron scanning microscope, and then use an atomic force microscope to test the same position to obtain mechanical and height information. This method has the disadvantages of low artificial experiment efficiency, low recognition accuracy and large error. [0004] Aiming at the problems of low identification accuracy and large errors in current mineralogical tests, a method for predicting rock...

AI Technical Summary

Problems solved by technology

[0003] The traditional mineralogy test method is to polish the tested rock material with argon ions, use the scanning electron microscope to obtain the imaging and shape information of the rock components, and then use the atomic force microscope to test the same position to obtain the mechanical and height information. This method has the disadvantages of low manual experiment efficiency, low recognition accuracy, and large errors.

[0004] Aiming at the problems of low recognition accuracy and large errors in current mineralogy testing, a method for predicting rock composition and mechanical properties is urgently needed

Images