DFPF-Net: Dynamically Focused Progressive Fusion Network for Remote Sensing Change Detection

TL;DR

This paper introduces DFPF-Net, a novel deep learning model combining transformers and CNNs with attention and edge detection to improve change detection accuracy in remote sensing images, effectively reducing noise and capturing true changes.

Contribution

The paper presents a new fusion network that integrates pyramid vision transformers with a dynamic focus module for enhanced change detection in remote sensing images.

Findings

Outperforms existing change detection methods on four datasets.

Effectively reduces noise caused by shadows and object variations.

Improves true change region detection accuracy.

Abstract

Change detection (CD) has extensive applications and is a crucial method for identifying and localizing target changes. In recent years, various CD methods represented by convolutional neural network (CNN) and transformer have achieved significant success in effectively detecting difference areas in bi-temporal remote sensing images. However, CNN still exhibit limitations in local feature extraction when confronted with pseudo changes caused by different object types across global scales. Although transformers can effectively detect true change regions due to their long-range dependencies, the shadows cast by buildings under varying lighting conditions can introduce localized noise in these areas. To address these challenges, we propose the dynamically focused progressive fusion network (DFPF-Net) to simultaneously tackle global and local noise influences. On one hand, we utilize a pyramid vision transformer (PVT) as a weight-shared siamese network to implement change detection, efficiently fusing multi-level features extracted from the pyramid structure through a residual based progressive enhanced fusion module (PEFM). On the other hand, we propose the dynamic change focus module (DCFM), which employs attention mechanisms and edge detection algorithms to mitigate noise interference across varying ranges. Extensive experiments on four datasets demonstrate that DFPF-Net outperforms mainstream CD methods.