A Multi-Source Data Fusion-based Semantic Segmentation Model for Relic Landslide Detection

TL;DR

This paper introduces HPCL-Net, a novel multi-source data fusion model using hyper-pixel contrastive learning for accurate relic landslide detection in remote sensing images, overcoming challenges like visual blur and limited data.

Contribution

The paper proposes a hyper-pixel-wise contrastive learning augmented segmentation network that fuses heterogeneous data sources and enhances feature extraction for landslide detection.

Findings

Significant improvement in mIoU from 0.620 to 0.651

Landslide IoU increased from 0.334 to 0.394

F1 score improved from 0.501 to 0.565

Abstract

As a natural disaster, landslide often brings tremendous losses to human lives, so it urgently demands reliable detection of landslide risks. When detecting relic landslides that present important information for landslide risk warning, problems such as visual blur and small-sized dataset cause great challenges when using remote sensing images. To extract accurate semantic features, a hyper-pixel-wise contrastive learning augmented segmentation network (HPCL-Net) is proposed, which augments the local salient feature extraction from boundaries of landslides through HPCL and fuses heterogeneous information in the semantic space from high-resolution remote sensing images and digital elevation model data. For full utilization of precious samples, a global hyper-pixel-wise sample pair queues-based contrastive learning method is developed, which includes the construction of global queues that store hyper-pixel-wise samples and the updating scheme of a momentum encoder, reliably enhancing the extraction ability of semantic features. The proposed HPCL-Net is evaluated on the Loess Plateau relic landslide dataset and experimental results verify that the proposed HPCL-Net greatly outperforms existing models, where the mIoU is increased from 0.620 to 0.651, the Landslide IoU is improved from 0.334 to 0.394 and the F1score is enhanced from 0.501 to 0.565.