The Tianshan Expressway plays a crucial role in China’s “Belt and Road” strategy, yet the extreme climate of the Tianshan Mountains poses significant traffic safety risks, hindering local economic development. Efficient detection of hazardous road surface conditions (RSCs) is vital to address these challenges. The complexity and variability of RSCs in the region, exacerbated by harsh weather, make traditional surveillance methods inadequate for real-time monitoring. To overcome these limitations, a vision-based artificial intelligence approach is urgently needed to ensure effective, real-time detection of dangerous RSCs in the Tianshan road network. This paper analyzes the primary structures and architectures of mainstream neural networks and explores their performance for RSC recognition through a comprehensive set of experiments, filling a research gap. Additionally, T-Net, specifically designed for the Tianshan Expressway engineering project, is built upon the optimal architecture identified in this study. Leveraging the split-transform-merge structure paradigm and asymmetric convolution, the model excels in capturing detailed information by learning features across multiple dimensions and perspectives. Furthermore, the integration of channel, spatial, and multi-head attention modules enhances the weighting of key features, making the T-Net particularly effective in recognizing the characteristics of snow-covered and icy road surfaces. All models presented in this paper were trained on a custom RSC dataset, compiled from various sources. Experimental results indicate that the T-Net outperforms fourteen once state-of-the-art (SOTA) models and three models specifically designed for RSC recognition, with 97.44% accuracy and 9.79% loss on the validation set.

The scarcity of road construction materials in arid desert areas necessitates the utilization of local resources to ensure sustainable infrastructure development. This study investigates the durability and damage evolution of a semi-rigid base material composed of Cement-Fly Ash Stabilized Gravel with 100% aeolian sand replacement for fine aggregates (CFSAG). The mechanical performance was evaluated under simulated desert environments, including high-temperature curing (30 °C, 40 °C, and 50 °C) and freeze–thaw cycles in both water and 2% Na₂SO₄ solution. Additionally, Digital Image Correlation (DIC) technology was employed to characterize the microscopic strain fields and crack propagation patterns. The results indicate that elevated curing temperatures significantly enhance compressive strength, with 40 °C identified as the optimal condition for strength formation due to accelerated hydration. Conversely, resistance to freeze–thaw cycles decreased with higher aeolian sand content, and sulfate erosion accelerated surface spalling, causing the specimen mass to exhibit an "increase-then-decrease" trend. A strength evolution model was established based on freeze–thaw frequency and mix proportions, achieving a fitting accuracy of over 98%. Furthermore, DIC revealed that the strain and displacement variations during crack propagation in specimens exhibited a three-stage evolutionary pattern, with strain localization bands becoming more pronounced as compaction decreases. The study concludes that CFSAG is a feasible, durable, and eco-friendly solution for semi-rigid pavement bases in desert regions.

Aeolian sand is abundant in arid deserts, but its high replacement in cement-stabilized bases can reduce strength and raise cracking risk. Strain localization and crack evolution are also poorly quantified. This study aimed to optimize the early age performance of cement-fly ash stabilized aeolian sand gravel (CFSAG) and clarify its failure mechanism. A Box–Behnken response surface methodology varied the cement content, cement-to-fly ash ratio, coarse aggregate gradation, and aeolian sand content. The 7-d unconfined compressive strength (UCS) and splitting tensile strength (STS) were tested. Digital image correlation (DIC) recorded full-field strains and crack metrics in compression and splitting. SEM–EDS was used to interpret microstructural changes. The aeolian sand content dominated UCS, whereas the cement content and cement-to-fly ash ratio mainly controlled STS. Factor interactions were non-negligible and supported the joint optimization of the two strength indices. DIC identified a crack propagation threshold near 0.9 P_max in splitting. Excess aeolian sand (>50%) caused earlier localization, more cracks, and wider openings. In the appropriate amount of aeolian sand mixtures, hydration products filled voids and improved paste continuity. SEM–EDS indicated that excessive fines increased porosity and weakened the interfacial transition zone. Overall, the combined RSM–DIC–SEM approach links mix design with deformation and microstructure evidence. It provides practical guidance to balance strength and cracking resistance at early ages for cement-stabilized bases in desert highway engineering.

In summary, near-surface icing risk in complex alpine terrain is jointly controlled by freezing conditions, moisture supply, freeze–thaw transitions, and topographic energy processes. Traditional approaches relying on sparse station data or single temperature thresholds fail to capture spatial heterogeneity, and frequent cloud cover together with topographic errors severely limit the application of thermal infrared remote sensing. Taking the area along the Duku Highway in the Tianshan Mountains as the study region, a daily icing risk assessment framework at 250 m resolution was constructed using multi-source remote sensing, ERA5-Land reanalysis data, topographic correction, and an energy–moisture dual-constrained model. A diurnal temperature cycle model, the CAP index, and physics-constrained machine learning were integrated to reconstruct the daily minimum land surface temperature ($T_{s,\mathrm{m}\mathrm{i}\mathrm{n}}$) at 250 m resolution under all weather conditions. A probabilistic two-tier risk assessment model was then established by incorporating moisture, topography, and freeze–thaw transitions. The results show that high-risk zones occur primarily in valleys and topographically constrained corridors rather than the coldest elevations. Validation against Landsat LST (r = 0.886) and the Bayanbulak station (bias −0.76 °C, RMSE 5.62 °C, r = 0.91) confirms spatial and seasonal accuracy. Sensitivity and Monte Carlo analyses indicate the RiskScore is mainly controlled by the low-temperature weight, while upstream parameters are less influential. The framework is best applied as a screening and early-warning product to identify sub-kilometer potential icing corridors, complementing point measurements and short-range forecasts.