Clay-bearing rocks slaking characterization using artificial vision

Several instability problems have been presented in cutting slopes over clay bearing rocks around the world. The principal cause is associated with the highly susceptibility to environmental changes of these materials. Their physical and mechanical characteristics change in a short period of time, thus, their performance change between rock and soil. This makes very complicated the classification process and the selection of suitable parameters for design and construction. Different classification proposals have been developed on fracture patterns basis and laboratory tests, but the subjectivity of the observer and the process for obtaining good samples affect the final result. Artificial vision is a good alternative for the characterization of these materials in the laboratory and the field. This paper investigates the disintegration of these materials using a new image segmentation algorithm. The algorithm works on color images, and different average color descriptors were used for the determination all factors which influence de segmentation quality.