Marine floating raft aquaculture extraction of hyperspectral remote sensing images based decision tree algorithm

The accurate extraction and mapping of floating raft aquaculture (FRA) is significant to the scientific management and sustainable development of coastal zones. However, the current relevant methods rely on large sample size and complex classifiers, which have poor generalization ability and thus are not suitable for large-scale application. To address these issues, this study proposes a new hyperspectral index based on remote sensing images, namely hyperspectral index for floating raft aquaculture (HSI-FRA). Based on the analysis of the spectral information, the HSI-FRA utilizes four bands at 580 nm, 740 nm, 1040 nm, and 1290 nm, respectively, to enhance the spectral difference between the floating raft aquaculture and the seawater through band calculation, and constructs the extraction index of the floating raft aquaculture. Using the HSI-FRA, a decision tree classification process is carried out to realize the extraction of floating raft aquaculture information. The extraction accuracy in the three study areas of Dayu Bay of Zhejiang Province, Fengwei Town of Fujian Province and Chao'an Bay of Guangdong Province is 94.3%, 95.5% and 91.44%, respectively, which has Larger advantage than those of traditional methods. The experimental results show that the proposed method is simple, fast and accurate at separating the floating raft aquacultures in an offshore complex marine environment.