A novel feature descriptor for automatic change detection in remote sensing images

Automatic change detection has expected increasing interest for researchers in recent years on high-spatial resolution remote sensing system where multispectral, multi-resolution and multimodal images can be acquired. The commonly used techniques for high-resolution change detection rely on feature extraction. Due to its high dimensional feature space, the conventional feature extraction techniques represent a progress of issues when handling huge size information e.g., computational cost, processing capacity and storage load. In order to overcome the existing drawback, we propose a novel Structural Phase Congruency Histogram (SPCH) descriptor for automatic change detection without reducing the significant loss of information. The proposed feature extractor depends upon the structural properties of the image which is invariant to contrast deviations and illumination. The structural phase congruency with the histograms is combined to build the edge and corner features. The dimensionality of the feature vector is reduced using Linear Discriminant Analysis (LDA) to form SPCH-LDA descriptor which leads to be more robust for image scale variations. Finally, the accuracy of the change detection is estimated with Artificial Neural Network (ANN) as compared with the existing algorithms. The experimental results provided 98.4375% accuracy which confirms the effectiveness and superiority of the proposed technique for automatic change detection. (C) 2018 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.