.png){kind=icon}
Return
Automatic processing method for SLR data based on dual-layer PSO-BP neural network
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asr.2026.08.035.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Current Satellite Laser Ranging (SLR) data preprocessing mostly relies on manual screen processing, which exhibits significant efficiency bottlenecks under large data throughputs and yields subjective inconsistencies. To address these challenges, this paper proposes an automatic SLR data processing method based on a dual-layer Particle Swarm Optimization (PSO) and Back Propagation (BP) neural network. This method effectively captures the local density fluctuations and covariance geometric features of SLR data. The outer-layer PSO adaptively searches for the optimal network structure, while the inner-layer PSO automatically optimizes the weights and biases. This dual-layer architecture mitigates the risk of the traditional BP algorithm getting trapped in local optima. Experimental results demonstrate that on simulated satellite data across different orbits, the method achieves a recognition rate exceeding 99.89% and a misdetection rate below 11.11% (remaining under 1.84% under standard low-noise conditions), outperforming both the hand-crafted Vanilla Neural Network (Vnn) and the classical Graz algorithm. For actual observation data, the internal precision of normal points is improved by up to 37.35% and 18.28% compared to ILRS official website reference data and the Graz algorithm, respectively. Additionally, this method supports batch processing with an average efficiency of approximately 25 to 32 s per file. Ultimately, it provides an efficient near-real-time automated preprocessing solution in the standard normal point generation pipeline for future unmanned SLR systems.
Keywords:
Satellite laser ranging
Data preprocessing
BP neural network
Dual-layer particle swarm optimization
Multi-scale features
Journal
IF:
2.8
Papers:
1.7K
Citations:
2.0W
Organization
Cited Papers
ShareSaveShareSaveShareSaveShareSaveShallow Water Bathymetry Mapping from ICESat-2 and Sentinel-2 Based on BP Neural Network Model.png)
WATER
OAAI
WATERIF3
OAAIShareSave2D Monte Carlo simulation of silicon waveguide-based single-photon avalanche diode for visible wavelengthsOPTICS EXPRESS
OAAI
OPTICS EXPRESSIF3.3
OAAIShareSave
View more