22 July 2022 in Geofluids
In the evaluation and prediction of slope stability, the traditional numerical analysis method, which is over reliant on experience, takes a large amount of computing time and lacks the ability to reflect the fuzzy and nonlinear characteristics of slope parameters well. Considering the above characteristics, this study proposes an improved particle swarm optimization of support vector machine (IPSO-SVM) algorithm model, which combines optimized particle swarm optimization (IPSO) and support vector machine (SVM) and applies it to slope stability prediction. Based on 28 groups of slope engineering data, the stability prediction results of IPSO-SVM, PSO-SVM, and SVM models were compared with real values for analysis. The results show that the maximum relative error of the IPSO-SVM model is only 1.3%, and the average relative error is 1.1%, which is far lower than the prediction error of the PSO-SVM model and SVM model; therefore, the prediction result of IPSO-SVM is the closest to the real value. This method can accurately predict the slope safety factor under the influence of different indexes, and the research results can provide guidance for practical engineering.
1 January 2021 in International Journal of Electrochemical Science
1 January 2025 in IET Image Processing
Low‐light images are prevalent in many real‐world scenarios, such as surveillance, autonomous driving, and mobile photography. However, these images often suffer from low contrast, high noise, and blurred details due to insufficient lighting. These degradations not only compromise visual aesthetics but also significantly hinder the performance of downstream tasks such as face recognition and security monitoring. To tackle these challenges, we propose the Fourier alignment network (FAN), which decomposes illumination and noise into amplitude and phase via frequency‐domain analysis and fuses spatial and frequency domain features in a cohesive framework. A spatial‐domain enhancement network effectively boosts illumination and suppresses artifacts, while gated convolution aligns illumination by capturing pixel‐wise lighting offsets. We further introduce focal frequency loss to minimize frequency‐domain discrepancies and complement traditional spatial losses, which guides the model to recover accurate illumination while preserving fine details. Comprehensive quantitative and qualitative experiments on the LOL, LOLv2, and UHDLL benchmarks have validated the effectiveness of our approach. Notably, on the UHDLL dataset, FAN achieves a PSNR of 26.28 dB and an SSIM of 0.92, which outperforms previous state‐of‐the‐art methods.