In this article, we develop two methods for unsupervised change detection in multitemporal remote sensing images based on Dempster-Shafer's theory of evidence (DST). In most unsupervised change detection methods, the probability of difference image is assumed to be characterized by mixture models, whose parameters are estimated by the expectation maximization (EM) method. However, the main drawback of the EM method is that it does not consider spatial contextual information, which may entail rather noisy detection results with numerous spurious alarms. To remedy this, we firstly develop an evidence theory based EM method (EEM) which incorporates spatial contextual information in EM by iteratively fusing the belief assignments of neighboring pixels to the central pixel. Secondly, an evidential labeling method in the sense of maximizing a posteriori probability (MAP) is proposed in order to further enhance the detection result. It first uses the parameters estimated by EEM to initialize the class labels of a difference image. Then it iteratively fuses class conditional information and spatial contextual information, and updates labels and class parameters. Finally it converges to a fixed state which gives the detection result. A simulated image set and two real remote sensing data sets are used to evaluate the two evidential change detection methods. Experimental results show that the new evidential methods are comparable to other prevalent methods in terms of total error rate.
Dynamic texture (DT) is an extension of texture to the temporal domain. Recognizing DTs has received increasing attention. Volume local binary pattern (VLBP) is the most widely used descriptor for DTs. However, it is time consuming to recognize DTs using VLBP due to the large scale of data and the high dimensionality of the descriptor itself. In this paper, we propose a new operator called orthogonal combination of VLBP (OC-VLBP) for DT recognition. The original VLBP is decomposed both longitudinally and latitudinally, and then combined to constitute the OC-VLBP operator, so that the dimensionality of the original VLBP descriptor is lowered. The experimental results show that the proposed operator significantly reduces the computational costs of recognizing DTs without much loss in recognizing accuracy.