The use of the image color information beyond color histograms has been limited for image retrieval. A reason is the lack of an accepted core basic color operation set on color images. With the grown interest in image retrieval applied to color images, new operators have been recently developed, having interesting properties. Opening distributions on images based on granulometries constitute an extremely useful tool in morphological tasks. Efficient techniques have been proposed for binary and grayscale images using linear openings. The present study extends the granulometry concept for color images. In addition, it addresses the development of a new morphological approach grounded on particle size distributions for color images and their use as an additional textural information to build queries over an image database.