Abstract: Rock thin section image is one of the most direct and effective means to identify shale laminae. Using image color segmentation to identify shale laminae is a conventional method.When it is applied to the marine-continental transitional shale with complex and diverse laminae morphology, the identification effect depends on the local features of the image and is greatly affected by the laminae morphology. To address the above problems, we propose a method to identify marine-continental transitional shale laminae structure by converting rock slice thin images into frequency domain images using two-dimensional discrete Fourier transform, extracting frequency domain image features with principal component analysis technology, and establishing characterization parameters of shale laminae structure development degree. The method was applied to the analysis of rock thin section images of the target reservoir in the study area, and the application results showed that: The method is more applicable to the complex and diverse marine-continental transitional shale shale strata than the conventional method, and the conformation rate of laminae identification reaches more than 90%. The method can provide strong support for shale structure analysis and anisotropy evaluation.

Key words: marine-continental transitional shale, laminae identification, rock thin section, image processing, Fourier transform, principal component analysis