A novel technique for segmentation of color images is proposed. The technique implements a thresholding approach based on the analysis of the hue histogram: a new function for detecting valleys of the histogram has been devised and tested. A new blurring algorithm for noise reduction that works effectively when used over hue image, has been also developed. A feed-forward neural network that learns to recognize the hue ranges of meaningful objects completes the segmentation process.Experimental results show that the proposed technique is reliable and robust even in presence of changing environmental conditions. Extended experimentation has been carried in the framework of the Robot Soccer World Cup Initiative (RoboCup). The approach is fully general and may be successfully employed in any intermediate level image-processing task, where the color is a meaningful descriptor.