Knowledge of the precise water depth in shallow areas of the ocean is of great significance to the safe navigation of ships and hydrographic surveying. Compared with traditional bathymetry, satellite remote sensing for water depth determination makes it possible to cover large areas by dynamic observation. In this paper, we conducted an optically shallow water bathymetric inversion study using a Stumpf empirical model, random forest model, neural network model, and support vector machine model based on Sentinel-2 satellite images and Ganquan Dao measured bathymetry data. We compared and analyzed the inversion results based on the empirical model and different machine learning models. The results show that the Stumpf empirical and machine learning models are capable of inverting optically shallow water depth. Moreover, the machine learning models had better fitting ability than the Stumpf empirical model with a sufficient number of samples, especially when the water depth was greater than 15 m. In addition, the random forest model had the highest overall accuracy among these models, with a root mean square error (RMSE) of 1.41 m and a regression coefficient (R²) of 0.96 for the test data.

The rapid expansion of aquaculture in coastal areas is typically associated with ecological negligence and low water quality owing to the economic exploitation of these areas. However, evaluation of these water bodies tends to be laborious, time-consuming, and costly. Therefore, to overcome the limitations of field surveys, in this study, we evaluated the water quality of the cultured water in the Beibu Gulf of Guangxi, obtained spectral reflectance by unmanned aerial vehicle with multispectral sensors, and constructed inverse models of 11 water quality parameters, namely, ammonia nitrogen (NH₃-N), chemical oxygen demand (COD), active phosphate (PO₄⁻), dissolved oxygen, nitrate nitrogen (NO₃-N), nitrite nitrogen (NO₂-N), inorganic nitrogen, total nitrogen, total phosphorus, suspended solids (SS), and chlorophyll a (chl-a), based on the partial least squares method to invert the water quality distribution of regional aquaculture. Furthermore, we compared the retrieval accuracy of different water quality parameters. The following results were obtained: 1) the constructed model’s results showed that the retrieval models for COD, NO₃-N, SS, and chl-a had better accuracy compared with those of other parameters; 2) application of the model to the validation set data yielded a correlation coefficient of 0.93 between the measured and predicted SS values, with a mean absolute error of prediction of 4.65 mg L⁻¹; this parameter constructed the best prediction model. According to the validation set results, the correlation coefficients of chl-a, COD and NO₃-N are all greater than 0.8, which had better performance effects compared with the remaining models, which are 0.87, 0.86, and 0.81 respectively. This study provides a reference for remote sensing monitoring of water quality in mariculture in cloudy and rainy areas.