To investigate the impact of image resolution and bathymetry models on the fusion of heterogeneous satellite data for depth inversion,this study integrates ICESat-2 laser altimetry data with multi-temporal Landsat 8,Sentinel-2,and WorldView-3 satellite data. Depth inversion is performed using logarithmic ratio model,multi-band model,BP neural network,support vector machine,random forest,and extreme gradient boosting. Experimental results show that the spatial resolution of the images has an insignificant effect on the accuracy of depth inversion. Considering both the accuracy and resolution of the inversion results,Sentinel-2 satellite data performs the best. Moreover,the XGBoost model outperforms other models in terms of inversion performance,achieving an optimal RMSE of 0. 51 meters in the Dongsha Atoll experimental area. These results provide valuable reference for coastal depth measurement based on the fusion of heterogeneous remote sensing satellite data.
关键词
水深反演/多光谱卫星遥感影像/冰、云和陆地高程卫星/数据融合/对比分析
Key words
water depth retrieval/multispectral satellite remote sensing imagery/ICESat-2/data fusion/contrastive analysis
维普
万方数据