Using computer-based digital image processing technology to conduct micro-description and quantitative characterization of rock features forms the foundation of geological data digitization. While rock image analysis under microscopes and with higher precision has progressed rapidly, image processing for field outcrops has received little attention. This paper takes clastic rocks as an example and explains techniques for processing petrological images from field outcrops. To address the issue of oversimplified quantitative representation of rock features, methods are explored in three areas: image denoising, morphological processing, and image segmentation. By linking the lithology of clastic rocks to their colors, a "penalty and reward scoring mechanism" algorithm is introduced, and 14 lithology classification situations based on color are summarized. A region-growing algorithm is also applied to accurately identify special minerals such as iron and brass. Edge detection, watershed algorithms, and other morphological techniques are then performed on the field images, followed by color-based particle labeling and calculation of related morphological parameters. The Folk-Ward formula is improved to analyze grain size results. Fractures and pores are quantitatively characterized next. Fractures are extracted automatically using morphological processing and adaptive binarization, with interactive computation of fracture lengths. The maximum internal tangent circle algorithm calculates the maximum fracture width. Connected regions are filtered based on color, size, and solidity to extract pores, allowing the area, perimeter, and circularity to be computed. The experimental results demonstrate that systematic use of computer image processing enables multi-dimensional feature extraction of field clastic rocks, thereby improving the efficiency of geological fieldwork.  Keywords ---Rock physics, Theory, Data processing, Computing aspects, Interpretation