Change detection in high-resolution remote sensing imagery is essential for a wide range of applications, including urban sprawl monitoring, conducting environmental assessments, and responding to disasters. Despite significant advancements in deep learning, challenges remain in capturing subtle changes, preserving spatial detail, and minimizing false detections. This work proposes a novel change detection framework, in which a high-resolution feature extraction backbone is integrated with dual attention mechanisms and multiscale contextual aggregation. In particular, HRNet serves as the backbone to obtain an informative representation of highr-esolution images, while both channel-wise and spatial attention modules are incorporated to enhance the representation’s discriminative capability. A residual change decoder jointly encodes absolute feature differences and semantic content, while a pyramid pooling module captures contextual dependencies across multiple scales. Finally, a lightweight refinement block is introduced to improve boundary sharpness and reduce noise. Extensive experiments on the LEVIR-CD dataset demonstrate that the proposed method achieves superior performance compared to state-of-the-art baselines, with improvements observed across major evaluation metrics. The obtained accuracy of the proposed model (F1-score of 89.39% and IoU of 80.82%) substantiates the robustness and effectiveness of the proposed architecture for reliable change detection in high-resolution remote sensing imagery.