Seasonally frozen soils are exposed to freeze‒thaw cycles every year, leading to mechanical property deterioration. To reasonably describe the deterioration of soil under different conditions, machine learning (ML) technology is used to establish a prediction model for soil static strength. Six key influencing factors (moisture content, compaction degree, confining pressure, freezing temperature, number of freeze‒thaw cycles and thawing duration) are included in the modelling database. The accuracy of three typical ML algorithms (support vector machine (SVM), random forest (RF) and artificial neural network (ANN)) is compared. The results show that the ANN outperforms the SVM and RF. Principal component analysis (PCA) is combined with the ANN, and the PCA–ANN algorithm is proposed, which further improves the prediction accuracy. The deterioration of soil static strength is systematically researched using the PCA–ANN algorithm. The results show that the soil static strength decreased considerably after the first several freeze‒thaw cycles before the strength plateau occurred, and the strength reduction increased significantly with increasing moisture content and compaction degree. The PCA–ANN model can generate a reasonable prediction for the static strength or other soil properties of seasonally frozen soil, which will provide a scientific reference for practical engineering.

Accurate identification of natural and mine earthquakes in mining areas is of great significance to the construction of secondary disaster warning networks. Based on 490 records of natural and mine earthquakes in the Hegang area from 2006 to 2017, this paper compares and analyzes the ground motion characteristics of the research samples (150 earthquake records and 200 mine earthquake records) and selects the key identification parameters of dominant frequency, Pm/Tc, and Sm/Tc. The correct identification rate of the test samples (60 seismic records and 80 mine earthquake records) is 95.7%, 91.4%, and 93.6%, respectively, and the actual threat rate is 90.8%, 83.3%, and 86.3%, respectively. Finally, based on the selected key identification parameters, a “three-parameter comprehensive gradient discriminant method” is proposed. The correct identification rate and actual threat rate are 99.3% and 98.4%, respectively, which can basically accurately identify natural and mine earthquakes. It provides a certain method and theoretical support for the mining area vibration identification method, safety production, and disaster warning.

Precise and intelligent crack detection is essential for prolonging the service life and ensuring cost-effective maintenance of pavement. This study develops a crack detection system inspired by computer vision, utilizing the attention mechanism and dense connectivity features. The coding component consists of ResNet-50 integrated with an attention mechanism, which improves feature extraction capabilities and rapidly identifies crack locations. During decoding, the characteristics of the encoding part and its low-level information are taken into account, enhancing segmentation accuracy. In comparison to other common approaches, it demonstrates superior accuracy in F1-Score, IoU, PA, and other evaluation metrics on the cracks dataset captured by the authors and CFD. The segmentation facilitates the automatic extraction of crack length, area, and maximum width by the linear fitting or point-by-point curve-fitting technique, grid method, and maximum inscribed circle approach, respectively. The results align well with the measured values, and the computational viability of the pavement condition index is analyzed, demonstrating its practicality in the intelligent assessment of pavement health.