SCDNET: A novel convolutional network for semantic change detection in high resolution optical remote sensing imagery

Daifeng Peng,Lorenzo Bruzzone,Yongjun Zhang,Haiyan Guan,Pengfei He

DOI: https://doi.org/10.1016/j.jag.2021.102465

IF: 7.5

2021-12-01

International Journal of Applied Earth Observation and Geoinformation

Abstract:With the continuing improvement of remote-sensing (RS) sensors, it is crucial to monitor Earth surface changes at fine scale and in great detail. Thus, semantic change detection (SCD), which is capable of locating and identifying “from-to” change information simultaneously, is gaining growing attention in RS community. However, due to the limitation of large-scale SCD datasets, most existing SCD methods are focused on scene-level changes, where semantic change maps are generated with only coarse boundary or scarce category information. To address this issue, we propose a novel convolutional network for large-scale SCD (SCDNet). It is based on a Siamese UNet architecture, which consists of two encoders and two decoders with shared weights. First, multi-temporal images are given as input to the encoders to extract multi-scale deep representations. A multi-scale atrous convolution (MAC) unit is inserted at the end of the encoders to enlarge the receptive field as well as capturing multi-scale information. Then, difference feature maps are generated for each scale, which are combined with feature maps from the encoders to serve as inputs for the decoders. Attention mechanism and deep supervision strategy are further introduced to improve network performance. Finally, we utilize softmax layer to produce a semantic change map for each time image. Extensive experiments are carried out on two large-scale high-resolution SCD datasets, which demonstrates the effectiveness and superiority of the proposed method.

remote sensing

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper aims to address the issue of Semantic Change Detection (SCD) in high-resolution remote sensing images. Specifically, the paper proposes a new convolutional network (SCDNet) for generating pixel-level semantic change maps on large-scale remote sensing datasets.

#### Main Issues
1. **Limitations of Traditional Methods**: Most existing SCD methods focus primarily on scene-level changes, resulting in semantic change maps with only coarse boundaries or sparse category information.
2. **Lack of Datasets**: Due to the scarcity of large-scale SCD datasets, existing methods lack precision and category information.

#### Solutions
1. **Proposing SCDNet**: Based on the Siamese UNet architecture, it includes two encoders and decoders that share weights to effectively combine multi-temporal image information.
2. **Multi-Scale Atrous Convolution (MAC)**: Inserting the MAC module at the end of the encoder to expand the receptive field and capture multi-scale information.
3. **Attention Mechanism and Deep Supervision Strategy**: Introducing attention mechanisms and deep supervision strategies to improve feature fusion and avoid gradient vanishing problems.
4. **Novel Loss Function**: Defining an inter-class loss function that combines Dice loss and Focal loss to address the issue of class imbalance.

### Summary
This paper addresses the problem of fine semantic change detection on large-scale remote sensing datasets by proposing SCDNet, which improves the accuracy and detail of change maps.