Unsupervised Object-Based Change Detection Via a Weibull Mixture Model-Based Binarization for High-Resolution Remote Sensing Images

Tianjun Wu,Jiancheng Luo,Jianwu Fang,Jianghong Ma,Xueli Song

DOI: https://doi.org/10.1109/lgrs.2017.2773118

IF: 5.343

2017-01-01

IEEE Geoscience and Remote Sensing Letters

Abstract:Object-based change detection (CD) is an effective method of identifying detailed changes in land features by contrastively observing the same areas of high-resolution remote sensing images at different times. Binarization is the important step in partitioning changed and unchanged classes in the unsupervised domain. We formulate a novel binarization technique based on the Weibull mixture model, where generated similarity measure images are modeled using a mixture of nonnormal Weibull distributions. The parameters in the model are further globally estimated by employing a genetic algorithm. Two data sets with high-resolution remote sensing images are used to evaluate the effectiveness of the proposed method. Experimental results demonstrate that the method allows better and more robust unsupervised object-based CD than do state-of-the-art threshold-based and clustering-based methods. Advantages of the proposed method are embodied in the modeling of relatively few data of the changed class with a skewed and long tail distribution.