WHDA-FCM: Wolf Hunting-Based Dragonfly With Fuzzy C-Mean Clustering For Change Detection In SAR Images

摘要

For the past few years, the automated addressing of changes in remote sensing images plays a significant role. However, the change detection (CD) model often suffers from the issue of speckle noise. More investigations have been proceeded to overcome this obstacle. This paper also considers the same issue and proposes a new CD model in synthetic aperture radar (SAR) images. Here, two SAR images that are captivated at different times will be considered as the input of the detection process. At first, discrete wavelet transform is incurred for image fusion, where the coefficients are optimally selected through a hybrid model that hybridizes the gray wolf optimization and dragonfly (DA) optimization. At last, the fused images after inverse transform are clustered via the fuzzy c-mean (FCM) clustering approach, and a similarity measure is performed between the segmented image and the ground truth image. The proposed model, wolf hunting-based DA with FCM, compares its performance over other conventional methods in terms of measures like accuracy, specificity, sensitivity, precision, negative predictive value, F_1 score and Matthews correlation coefficient. Similarly, the negative measures are false positive rate, false negative rate and false discovery rate, and the betterment is proven.