Landslide Detection of Hyperspectral Remote Sensing Data Based on Deep Learning With Constrains

Chengming Ye, Yao Li, Peng Cui, Liang Li, José Marcato, Wesley Nunes Gonçalves, Jonathan Li

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · 2019 · 163 citations · 90 references

Concepts

TL;DR

Detecting and monitoring landslides is a hot topic in remote sensing, yet no study has applied deep learning to hyperspectral images for this task. The study proposes a constraint‑based deep learning framework to detect landslides in hyperspectral images. The framework consists of two steps: a deep belief network extracts spectral–spatial features, and a logistic regression classifier verifies landslides using high‑level features and constraints. Experimental results show the framework achieves higher overall accuracy than traditional hyperspectral classification methods, with a precision of 97.91% versus 94.36%, 84.50%, and 86.44% for linear SVM, spectral information divergence, and spectral angle match, respectively, and demonstrate the high‑level feature extraction system’s potential for multi‑source remote sensing.

Abstract

Detecting and monitoring landslides are hot topics in remote sensing community, particularly with the development of remote sensing technologies and the significant progress of computer vision. To the best of our knowledge, no study focused on deep learning-based methods for landslide detection on hyperspectral images. We proposes a deep learning framework with constraints to detect landslides on hyperspectral image. The framework consists of two steps. First, a deep belief network is employed to extract the spectral–spatial features of a landslide. Second, we insert the high-level features and constraints into a logistic regression classifier for verifying the landslide. Experimental results demonstrated that the framework can achieve higher overall accuracy when compared to traditional hyperspectral image classification methods. The precision of the landslide detection on the whole image, obtained by the proposed method, can reach 97.91%, whereas the precision of the linear support vector machine, spectral information divergence, and spectral angle match are 94.36%, 84.50%, and 86.44%, respectively. Also, this article reveals that the high-level feature extraction system has a significant potential for landslide detection, especially in multi-source remote sensing.

References

90