Enteromorpha Prolifera Detection in High-Resolution Remote Sensing Imagery Based on Boundary-Assisted Dual-Path Convolutional Neural Networks

Enteromorpha prolifera is a frequent marine ecological environment disaster. How to quickly and accurately monitor E. prolifera is of great significance to its management and protection of the marine ecological environment. The detection of E. prolifera from high spatial resolution remote sensing images (HSRIs) is an important technical means for monitoring E. prolifera disasters. With respect to the difficulty in accurate detection of E. prolifera area boundary in HSRIs, this article proposes an E. prolifera detection method for HSRIs based on boundary-assisted dual-path convolutional neural networks (BADP-CNNs). First, a large-scale HSRIs' E. prolifera detection dataset, FIO-EP, is created and published to facilitate the field of HSRIs' E. prolifera detection. Then, a BADP-CNN framework is designed to detect E. prolifera in HSRIs based on the shape distribution characteristics of E. prolifera. In the CNN framework, accurate detection of E. prolifera areas in HSRIs is achieved by fusing initial detection and boundary detection results of E. prolifera. The proposed method is compared with some state-of-the-art E. prolifera detection algorithms using the FIO-EP dataset. The experimental findings demonstrate that the proposed method can obtain 88.28% F1-score and 79.02% intersection-over-union (IOU) and is superior to other state-of-the-art E. prolifera detection algorithms.