Improved and Optimized GNSS-IR Sea Surface Height Retrieval Based on Noise Elimination and Lightweight Airborne Multi-GNSS Multi-UAV Fusion

Global Navigation Satellite System (GNSS) reflection remote sensing technology enriches the existing earth observation technologies. It is urgent to carry out high-precision inversion research of GNSS-IR sea surface height on UAV telemetry platforms equipped with GNSS receivers and lightweight receiver antennas. This study focuses on two aspects of work in DJI Mavic 3 Enterprise (DJI M3E). On the one hand, taking into account the extraction of Signal to Noise Ratio (SNR) data and the separation of direct signal and reflected signal components, an improved SNR residual (δSNR) data noise elimination method based on Variational Mode Decomposition for Marine Predator Algorithm Optimization (MPA-VMD) algorithm is proposed. The inversion accuracy of the improved model is better, which confirms the effectiveness of the improved GNSS-IR model based on intelligent extraction, decomposition and reconstruction of δSNR data. On the other hand, a method of airborne multi-GNSS and multi-UAV collaborative fusion based on truncated mean optimization is established according to the multi-frequency and multi-SNR type. It can be seen that the experimental scheme of multi-UAV fusion first and then multi-GNSS fusion has the highest accuracy, with a reflection height of 9.05 m and an error of only 0.27 m. Compared with before optimization, RMSE increased by 3.89% and ME decreased by 2.17%, verifying the reliability and accuracy of the GNSS-IR optimization model based on airborne multi-GNSS and multi-UAV collaborative fusion. In summary, the improved model based on noise elimination and the optimized model of airborne multi-GNSS multi-UAV collaborative fusion can obtain robust, reliable results.