Abstract: A color feature extraction method based on HSI(Hue-Saturation-Intensity) color components is proposed.The presented method integrates the characteristics that HSI color components express the intrinsic color information with the advantage that the multi-threshold classification method of component histogram adapts to the image content,quantifies HSI color components using multi-threshold classification method based on histogram,and extracts image color features by combining quantified color components.Color features of visual image extracted by this method are clustered,the visual image is segmented,and the segmentation results prove that the color feature extraction method is effective and feasible.For the excessive classification problem of color components caused by local maximums of color component histogram,a method is presented to eliminate local maximums in the color component histogram by amending the data between maximum pairs.Data distribution characteristics of the color component histogram are analyzed quantificationally,and then the local maximums in color component histogram are confirmed and amended.The experiment results indicate that the presented method can effectively eliminate local maximums in the histogram data.