NFANet: A Novel Method for Weakly Supervised Water Extraction from High-Resolution Remote Sensing Imagery

TL;DR

NFANet is a novel weakly supervised deep learning method that effectively extracts water bodies from high-resolution remote sensing images using point labels, neighbor feature aggregation, and recursive training.

Contribution

The paper introduces NFANet, a new approach leveraging neighbor feature aggregation and recursive training to improve water extraction accuracy with point labels.

Findings

Outperforms other weakly supervised methods

Achieves similar accuracy to state-of-the-art fully supervised methods

Effectively captures natural water boundaries

Abstract

The use of deep learning for water extraction requires precise pixel-level labels. However, it is very difficult to label high-resolution remote sensing images at the pixel level. Therefore, we study how to utilize point labels to extract water bodies and propose a novel method called the neighbor feature aggregation network (NFANet). Compared with pixellevel labels, point labels are much easier to obtain, but they will lose much information. In this paper, we take advantage of the similarity between the adjacent pixels of a local water-body, and propose a neighbor sampler to resample remote sensing images. Then, the sampled images are sent to the network for feature aggregation. In addition, we use an improved recursive training algorithm to further improve the extraction accuracy, making the water boundary more natural. Furthermore, our method utilizes neighboring features instead of global or local features to learn more representative features. The experimental results show that the proposed NFANet method not only outperforms other studied weakly supervised approaches, but also obtains similar results as the state-of-the-art ones.