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

As a new method of sea surface height monitoring, Global Navigation Satellite System (GNSS) reflection remote sensing technology enriches the existing earth observation technologies and provides a large number of ground reflection information sources. Many scholars are focusing on ground-based research. Facing the increasingly rich demand for ocean surveying and mapping, 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. Based on this, this study focuses on two aspects of work in DJI Mavic 3 Enterprise (DJI M3E). On the one hand, at the level of airborne GNSS observation data, 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. Comparing the inversion accuracy before and after the improvement, it can be found that the inversion accuracy of the improved model is better as a whole, which confirms the effectiveness of the improved GNSS-IR model based on intelligent extraction, decomposition and reconstruction of δSNR data; Moreover, the reflection heights of Drone No.3, Drone No.4 and Drone No.5 under different satellites are not much different, which further proves the stability of the model. On the other hand, based on the GNSS-IR improvement model, 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 accuracy analysis and evaluation results. From the experimental results, 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, under the background of GNSS reflection remote sensing technology provided in this study, the improved model based on noise elimination and the optimized model of airborne multi-GNSS multi-UAV collaborative fusion can obtain robust, reliable, and high-precision sea surface height monitoring results of the UAV telemetry platform, providing an important reference for high-precision sea surface height inversion in offshore and open seas.