It is essential to quantitatively evaluate road safety compliance through the modeling of three-dimensional geometric information of in-service roads to automatically guide the determination of driving speeds in intelligent transportation. A novel methodology is presented for reliably estimating 3D geometric elements and assessing geometry-related safety. A point-octree-normal-vector (PONV) network is developed to cope with the heterogeneous density of point clouds from mobile and terrestrial laser scanning and reduce possible failures caused by the disordered rotation of points. The points on road surfaces are adaptively segmented into stripes without data loss. Road horizontal elements, cross-sectional slopes and longitudinal slopes are estimated independent of alignment variations based on a concave hull algorithm, a cubic B-spline algorithm and a total least-squares method. A road safety compliance assessment related to the radius of curvature, average gradient, rate of curvature changes and side friction factors is conducted. Warning and suggested driving speeds are thus provided. Experiments show that the PONV network achieved superior performance in extracting road surfaces, with accuracy, precision, recall, and mean intersection-over-union scores of 98.1%, 89.7%, 89.3% and 82.5%, respectively. The geometric elements with neighboring distances in meters were basically consistent with those in traditional surveys and raw point clouds. Through comparison with safety guidelines, suitable vehicle operation speeds were suggested in road sections with unreasonable rates of curvature changes and side friction factors. This work can be applied for constructing road information models and automatically guiding driving speeds with high-definition maps on autonomous vehicles.

Classifying point clouds obtained from mobile laser scanning of road environments is a fundamental yet challenging problem for road asset management and unmanned vehicle navigation. Deep learning networks need no prior knowledge to classify multiple objects, but often generate a certain amount of false predictions. However, traditional clustering methods often involve leveraging a priori knowledge, but may lack generalisability compared to deep learning networks. This paper presents a classification method that coarsely classifies multiple objects of road infrastructure with a symmetric ensemble point (SEP) network and then refines the results with a Euclidean cluster extraction (ECE) algorithm. The SEP network applies a symmetric function to capture relevant structural features at different scales and select optimal sub-samples using an ensemble method. The ECE subsequently adjusts points that have been predicted incorrectly by the first step. The experimental results indicate that this method effectively extracts six types of road infrastructure elements: road surfaces, buildings, walls, traffic signs, trees and streetlights. The overall accuracy of the SEP-ECE method improves by 3.97% with respect to PointNet. The achieved average classification accuracy is approximately $99.74\%$ , which is suitable for practical use in transportation network management.

In spaceborne microwave remote sensing and geodesy, tropospheric delay has emerged as a critical factor affecting the precision of measurements. While the Global Navigation Satellite System (GNSS) offers reliable station-wise Zenith Total Delay (ZTD) products, their spatial resolution is inherently constrained by the GNSS stations distribution. Conversely, empirical models combined with numerical weather models (NWM), such as the fifth generation of European Reanalysis (ERA5) and Vienna Mapping Functions 3 (VMF3), can generate global gridded ZTD estimates. Yet, they exhibit centimeter-level discrepancies when benchmarked against GNSS-derived ZTD. This paper proposes a deep learning method based on the Gaussian Mixture Long Short-Term Memory (GM-LSTM) network, which learns the mapping of probability density between ZTD derived from the empirical model to the ZTD derived from GNSS. Once this mapping is learned, it can be used to infer the ZTD probability distribution and its uncertainty at any location within the study area. Upon evaluation across eight different latitude regions in Europe, the ZTD inferred by the proposed GM-LSTM model reaches the state-of-the-art level with an average Root Mean Square Error (RMSE) of 4.6mm. Compared with ZTD estimated from Deep Neural Network (DNN), ERA5 ray tracing, VMF3, and the Generic Atmospheric Correction Online Service (GACOS), the proposed GM-LSTM model achieved average performance improvements of 41.78%, 68.20%, 49.56%, and 50.43%, respectively. Verified by the meteorological records, the proposed GM-LSTM model can effectively reflect the uncertainty caused by spatially heterogeneous rainfall events. With homogeneous training data, it shows good performance in heavy rainfall, which is not matched by other ZTD estimation methods.

Water vapor, the most influential greenhouse gas, is central to Earth’s climate system, affecting the hydrological cycle, energy balance, and atmospheric dynamics. Integrated Water Vapor (IWV) is a key variable for understanding these processes. However, conventional IWV retrieval methods—such as ground-based sensors, satellite observations, and numerical weather models (NWM)—are often limited by spatial resolution, temporal continuity, and retrieval accuracy. To address these challenges, this study introduces a novel deep learning method GMLSTM-HIM, a High-resolution IWV estimation Model (HIM) based on a Gaussian Mixture Long Short-Term Memory (GMLSTM) framework. By integrating Global Navigation Satellite System (GNSS) and NWM inputs, including weighted mean temperature, GMLSTM-HIM utilizes a bidirectional LSTM structure and probabilistic output sequences to improve IWV estimation accuracy while quantifying uncertainty arising from spatial heterogeneity. Compared to ERA5 and VMF3, the model achieves average RMSE reductions of 68.44% and 36.15%, respectively. The model’s performance is further evaluated through inter-comparisons with MODIS and Fengyun satellite-derived IWV products, highlighting both the accuracy of GMLSTM-HIM and the complementary strengths of satellite observations. The results suggest that, of the satellite datasets examined in this case study, the MODIS 5 km product exhibits the highest consistency with the GMLSTM-HIM model estimates, outperforming the higher-resolution MODIS 1 km and FY-3D 1 km products in terms of product reliability (measured by root mean square error and correlation). A data fusion strategy is also proposed, combining model and satellite estimates to preserve fine-scale details and enhance robustness. Overall, GMLSTM-HIM provides a robust framework for high-resolution IWV retrieval, with significant potential to advance atmospheric studies, climate surveillance, and operational weather forecasting within the remote sensing community.

Water vapor, the most significant greenhouse gas, plays a critical role in the Earth's climate system, influencing the hydrological cycle, energy distribution, and atmospheric dynamics. Integrated Water Vapor (IWV) serves as a vital parameter for understanding these processes. While traditional methods, including ground-based instruments and satellite observations, provide IWV measurements, they are often limited by spatial resolution, coverage, and accuracy. Advances in numerical weather models (NWM) and remote sensing have improved large-scale IWV estimations but still face challenges in capturing high accuracy. To address these limitations, this study introduces a novel approach using a Gaussian Mixed Long Short-Term Memory (GM-LSTM) deep learning model to generate high-resolution water vapor fields (WVF) with enhanced spatial and temporal resolution. The GM-LSTM integrates numerical weather models (NWM) and GNSS data to create an adaptive mapping for zenith wet delay (ZWD) estimation, which is then converted to IWV. By utilizing a bidirectional Long Short-Term Memory (Bi-LSTM) architecture and probabilistic density distribution sequences, the model not only improves ZWD estimation accuracy but also quantifies inherent uncertainties due to spatial heterogeneity. Compared with ERA5 and VMF3, the proposed GM-LSTM achieves an average RMSE reduction of 67.68% and 48.74%. The proposed WVF generation method was validated through an inter-comparison with MODIS and Fengyun satellite products, highlighting its superior accuracy and reliability. This study demonstrates the potential of deep learning models like GM-LSTM to overcome the limitations of traditional techniques, providing a transformative tool for high-resolution IWV estimation and supporting advancements in climate monitoring and weather prediction.Keywords:WVF, Inter-comparison, IWV, GM-LSTM, GNSS, MODIS, Fengyun satellite

In recent years, geodesy based on spaceborne microwave remote sensing has gained significant advances. However, whether the observations from the Global Navigation Satellite System (GNSS) or Interferometric Synthetic Aperture Radar (InSAR), the results are inevitably influenced by atmospheric tropospheric delay. Although the tropospheric zenith total delay (ZTD) can be estimated through the gridded meteorological data products and empirical models provided by the ERA5 reanalysis product, its accuracy is still insufficient to meet the needs of modern geodesy. To overcome this challenge, we propose leveraging machine learning techniques to learn local spatio-temporal patterns of tropospheric delay for inferring total zenith delay (ZTD) and zenith wet delay (ZWD) at any location within the learning area.Our findings indicate that artificial neural networks can establish a robust mapping between ZTD estimated by empirical models and GNSS-measured ZTD. Then employing the ensemble learning strategy and the time series dynamics model, the ZTD at any location within the sample area can be inferred. To evaluate our approach, we conducted tests during the active water vapor season in the Tübingen region of Baden-Württemberg, Germany, from June 25 to July 9, 2022. In comparative experiments with the root mean square error (RMSE) of Zenith Total Delay (ZTD) derived from ERA5, our proposed method yielded a significant reduction in RMSE, decreasing it from 16.4292mm to 7.2108mm. This reflects a remarkable accuracy improvement of 56.11%.The proposed approach holds promise for enhancing the precision of GNSS positioning, InSAR earth observation, and generating more dependable water vapor products.