Abstract:Change detection in remote sensing images is essential for monitoring alterations in the Earth's surface over time. Despite numerous methodologies suggested for synthetic aperture radar (SAR) image classification and change detection, achieving both high accuracy and computational efficiency remains a challenge. To address this gap, present the proportional factor-based firefly optimization and multilayer perceptron deep neural network (PFFO-MPDNN) framework tailored for SAR image classification and change detection (SAR-IC-CD). This approach begins by harnessing data from diverse geographical locations, including Bern, Ottawa, Mexico, and the Yellow River datasets. employing bilateral and spatial filtering (BSF) for noise reduction and modified contrast-limited adaptive histogram equalization (MCLAHE) for contrast enhancement, prepare the data for subsequent analysis. Leveraging the modified gauss log ratio (MGLR) operator, derive dissimilar image representations, enriching the dataset's discriminative power. Next, extract features using ternary pattern and discrete wavelet transform (TP-DWT), including SPCH, gray-level co-occurrence matrix, run length, Shannon entropy, and Pearson correlation. Through proportional factor-based firefly optimization (PFFO), select optimal features to enhance classification accuracy. The selected features are then fed into a multilayer perceptron and deep neural network (MPDNN) for image classification, effectively distinguishing between changed and unchanged image parts. This experiment is conducted in MATLAB, and it demonstrates the superior performance of the proposed PFFO-MPDNN-SAR-IC-CD method across various performance metrics, including accuracy, sensitivity, specificity, rand index, global consistency error, variation of information, false-negative rate (FNR), and false-positive rate (FPR). Comparisons with existing methods, including RBF-DCNN-SAR-IC, DBN-SAR-IC, CNN-SAR-IC, USF-SAR-IC, DDN-SAR-IC, L-CapsNet-SAR-IC, SVM-HRLSM-SAR-IC, and Faster R-CNN-SAR-IC, demonstrate the superiority of this approach in SAR image classification and change detection tasks. This method achieves significant accuracy improvements, particularly in the Yellow River dataset, showcasing its effectiveness and robustness in real-world applications.

Change detection in SAR image based on weighted difference image generation and optimized random forest

SAR Image Change Detection Via Heterogeneous Graph with Multi-order and Multi-level Connections

Unsupervised SAR Image Change Detection Based on Structural Consistency and CFAR Threshold Estimation

Change Detection from SAR Images Based on Convolutional Neural Networks Guided by Saliency Enhancement

Change Detection in Multitemporal SAR Images Based on Slow Feature Analysis Combined with Improving Image Fusion Strategy

Automatic SAR Change Detection Based on Visual Saliency and Multi-Hierarchical Fuzzy Clustering

Change Detection in Multi-Temporal SAR Images Based on Slow Feature Analysis Combined With Improving Image Fusion Strategy

Synthetic Aperture Radar Image Change Detection Based on Principal Component Analysis and Two-Level Clustering

Enhanced Edge Information and Prototype Constrained Clustering for SAR Change Detection

Synthetic aperture radar image change detection using saliency detection and attention capsule network

Optimizing synthetic aperture radar image classification and change detection: a proportional factor firefly algorithm and multilayer perceptron approach

A Robust Imbalanced SAR Image Change Detection Approach Based on Deep Difference Image and PCANet

Synthetic Aperture Radar Image Change Detection via Siamese Adaptive Fusion Network

Improved Difference Images for Change Detection Classifiers in SAR Imagery Using Deep Learning

SAR Image Change Detection via Heterogeneous Graph With Multiorder and Multilevel Connections

Change Detection in SAR Images Based on the ROF Model Semi-Implicit Denoising Method

Multidirectional Attention Fusion Network for SAR Change Detection

A novel approach for object-based change image generation using multitemporal high-resolution SAR images

Two-Phase Object-Based Deep Learning for Multi-Temporal SAR Image Change Detection

Change Detection in SAR Images Based on Improved Non-Subsampled Shearlet Transform and Multi-Scale Feature Fusion CNN

Spaceborne SAR Time-Series Images Change Detection Based on SAR-SIFT-Logarithm Background Subtraction