SAR Image Change Detection Based on Multiscale Capsule Network
This work addresses change detection in SAR images, which is crucial for remote sensing applications, but it appears incremental as it builds on existing capsule network concepts with specific adaptations.
The paper tackled the problem of synthetic aperture radar image change detection by proposing a Multiscale Capsule Network to address challenges like speckle noise and deformation sensitivity, achieving improved robustness and efficiency as verified on three real SAR datasets compared to four state-of-the-art methods.
Traditional synthetic aperture radar image change detection methods based on convolutional neural networks (CNNs) face the challenges of speckle noise and deformation sensitivity. To mitigate these issues, we proposed a Multiscale Capsule Network (Ms-CapsNet) to extract the discriminative information between the changed and unchanged pixels. On the one hand, the multiscale capsule module is employed to exploit the spatial relationship of features. Therefore, equivariant properties can be achieved by aggregating the features from different positions. On the other hand, an adaptive fusion convolution (AFC) module is designed for the proposed Ms-CapsNet. Higher semantic features can be captured for the primary capsules. Feature extracted by the AFC module significantly improves the robustness to speckle noise. The effectiveness of the proposed Ms-CapsNet is verified on three real SAR datasets. The comparison experiments with four state-of-the-art methods demonstrate the efficiency of the proposed method. Our codes are available at https://github.com/summitgao/SAR_CD_MS_CapsNet .