Research On The Image Retrieval Technology Based On Color And Edge Features

In recent years, with the digital image number growing rapidly, intend to retrieve useful information fast and accurately, content-based image retrieval technology got demand and fast development. But fuzziness of image information brings difficulties for feature extraction technology. Therefore how to extract image feature accurately and comprehensively becomes the difficulty of the content-based image retrieval technology research.In the existing domestic and overseas scholars and researchers research fruits, most of the image retrieval technologies take the bottom features of image (color feature, texture feature, shape feature) as the basic for research. Based on the previous studies, this paper has researched and analyzed the bottom features of image deeply, and study the technology retrieving images using based on color and edge features. The technology uses improved HSV color histogram and mathematical morphology to obtain the similarity measurement vector of color feature and the similarity measurement vector of edge feature respectively, then, gives the simulation experiment respectively. Finally, combines the similarity measurement vector of color feature and the similarity measurement vector of edge feature with certain weights to retrieve images, and designs and builds retrieval system.Simulation experiments show that the image retrieval technology of weighting color and edge features overcome the insufficient of a single image feature retrieval of color and edge features. The weighting algorithm yields preferable results, and it has certain application value.