In order to alleviate the impact of radio frequency interference (RFI) on the accuracy of ocean salinity satellite remote sensing, scholars have proposed various detection and labeling algorithms for RFI based on remote sensing data from the SMOS satellite. However, the signals that generate RFI are diverse, and the factors that influence remote sensing observation data are complex. Existing algorithms often target specific hypothetical conditions, lacking general applicability, which frequently leads to an important gap between the nominal performance of the literature and practical applications, posing great challenges to data labeling work. To address this problem, this study conducted a comprehensive and systematic analysis of RFI simulation based on scene modeling, algorithm modeling, and RFI energy modeling. Three typical RFI detection algorithms were selected, and the simulation scene was divided into 3 typical scenes: ocean, land, and sea–land scenes, and RFI was analyzed in terms of weak, moderate, strong, and extremely strong based on energy. Through simulation analysis and evaluation of RFI detection algorithms, lookup tables for algorithm selection, detection rate, and false-positive rate have been established for different intensities of independent RFI sources and multiple nearby RFI sources in the above scenario. These lookup tables have universal guiding significance and provide reliability assurance in complex situations.

The distribution and change of sea surface salinity (SSS) have an important influence on the sea dynamic environment, marine ecological environment, global water cycle, and global climate change. Satellite remote sensing is the only practical way to continuously observe SSS over a wide area and for a long period of time. The salinity retrieval model of flat sea surface, which primarily includes empirical model and iterative model, is the key to retrieving satellite SSS products. The empirical models have high computational efficiency but low inversion accuracy, while the iterative models have high inversion accuracy but low computational efficiency. In order to reconcile the contradiction between the computational efficiency and inversion accuracy of existing models, this paper proposes a universal deep neural network (DNN) model architecture and corresponding training scheme, and provides 3 DNN models with extremely high computational efficiency and high inversion accuracy. The inversion error range, the root mean square error (RMSE), and the mean absolute error (MAE) of the DNN models on 311,121 sets of data have decreased by more than 40 times, 150 times, and 150 times, respectively, compared to the empirical model. The computational efficiency of the DNN models on 420,903 sets of data has improved by more than 100,000 times compared to the iterative model. Therefore, the algorithm developed in this paper can effectively solve the contradiction between the computational efficiency and inversion accuracy of existing models, and provide a theoretical support for high-precision and high-efficiency salinity inversion research.