Color and Texture Feature for Remote Sensing – Image Retrieval System: A Comparative Study

In this study, we proposed score fusion technique to improve the performance of remote sensing image retrieval system (RS-IRS) using combination of several features. The representation of each feature is selected based on their performance when used as single feature in RS-IRS. Those features are color moment using L*a*b* color space, edge direction histogram extracted from Saturation channel, GLCM and Gabor Wavelet represented using standard deviation, and local binary pattern using 8-neighborhood. The score fusion is performed by computing the value of image similarity between an image in the database and query, where the image similarity value is sum of all features similarity, where each of feature similarity has been divided by SVD value of feature similarity between all images in the database and query from related feature. The feature similarity is measured by histogram intersection for local binary pattern, whereas the color moment, edge direction histogram, GLCM, and Gabor are measured by Euclidean Distance. The final result shows that the best performance of remote sensing image retrieval in this study is a system which uses the combination of color and texture features (i.e. color moment, edge direction histogram, GLCM, Gabor wavelet, and local binary pattern) and uses score fusion in measuring the image similarity between query and images in the database. This system outperforms the other five individual feature with average precision rates 3%, 20%, 13%, 11%, and 9%, respectively, for color moment, edge direction histogram, GLCM, Gabor wavelet, and LBP. Moreover, this system also increase 17% compared to system without score fusion, simple-sum technique.