To achieve a correct colour image segmentation in the presence of shadows is still a challenge. The varying object shapes in a scene provoke several effects in an image such as shadows, shadings, and highlights. This physical effects produce strong edges on an image that are not considered for human segmentation. These edges are detected by the current segmentation methods. This paper analyze the capability of the Ridge-based Distribution Analysis method (RAD) to perform a correct segmentation in the presence of shadows. RAD is based on the idea that the representative colour of a shadow and the representative colour of the object affected by this shadow share some spatial properties. RAD finds this relationship assigning one colour for an object even when this object is affected by shadows. This analysis is done on different chromatic spaces to determine the one that fits better to RAD. Results achieved point out that RAD is able to perform a correct segmentation in the presence of self and cast shadows.