Habitat Prediction of Northwest Pacific Saury Based on Multi-Source Heterogeneous Remote Sensing Data Fusion

Accurate habitat prediction is important to improve fishing efficiency. Most of the current habitat-prediction studies use single-source datasets and sequence model based on datasets, which, a certain extent, limits further improvement accuracy. In this paper, we propose method multi-source heterogeneous remote-sensing data fusion, using product-level L1B-level original data. We designed feature extraction Convolution Neural Network (CNN) Long Short-Term Memory network (LSTM), decision-fusion extraction. for Northwest Pacific Saury, mean R2 reaches 0.9901 RMSE decreases 0.01588 in validation experiment. It significantly better than results other models, with single as input. Moreover, performs well generalization experiment because limited error less 8%. Compared existing literature, proposed paper solves problem ineffective fusion caused by differences structure size through multilevel decision it deeply explores features fishery different structures sizes. can effectively accuracy prediction, proving feasibility advancement prediction. also provides new methods ideas future research field