Quantitative evaluation of rock weathering using image processing

The purpose of this research is to automate geological surveys using imaging analysis. In this paper, we discuss an imaging analysis method for quantitative evaluation of weathering. We took pictures of core samples gathered from a boring hole with a color CCD camera. Subsequently, the relation between (I1, I2, I3) color space information in these images in the color space and the weathering characteristics was discussed. Imaging analysis revealed that Inter Quartile Range and Median are closely related to degree of rock weathering. Effective parameters depending on rock types were selected based on this information, and then the degree of rock weathering was evaluated with Support Vector Regression. Using this method degree of rock weathering was correctly identified within a range of ±0.5 from the true value in 90% in the various rock types studied. The above results confirmed the validity of the quantitative evaluation method, suggesting its usefulness in automation of geological surveys. © 2010 The Institute of Electrical Engineers of Japan.