Remote Sensing Geological Classification of Sea Islands and Reefs Based on Deeplabv3 +

Abstract

China’s vast sea area, coral reefs are important marine resources, but also an important part of land resources. Remote sensing monitoring of coral reefs and studying their geological conditions and changes are of great significance to the development of coral reefs and environmental protection. In recent years, with the promotion of high-resolution remote sensing images, high-resolution remote sensing images are widely used in the classification of coral reefs. Aiming at the problems of fuzzy boundary and low classification accuracy in traditional methods for coral reef classification, a classification method based on Deeplabv3 + network model is proposed. Seven geological types of Ganquan Island in Xisha Islands are extracted and classified. The overall classification accuracy and Kappa coefficient are 97.57 % and 0.9643, respectively. Compared with traditional support vector machine (SVM) and object-oriented classification methods, the results show that: the geological classification of coral reefs from high-resolution satellite images based on Deeplabv3 + model can better extract its essential characteristics, and the classification effect is good, meeting the accuracy requirements.