With the global shift in energy systems and the growing adoption of renewable energy, photovoltaic power generation has become widely implemented worldwide as a clean and efficient energy source. Therefore, scientific and rational locating methods for photovoltaic power stations has become essential. This study employs a particle swarm optimization (PSO) model based on the fuzzy analytic hierarchy process (FAHP) for evaluating the site-suitability of photovoltaic power stations. Taking Qujing City in Yunnan Province as an example, this study comprehensively considered 13 factors in three categories: meteorology, physical geography, and location. In addition, such topographic factors as planar curvature, profile curvature, and LSW were added to the physical geography factors for consideration. In previous studies, less attention has been paid to these factors. Spatial analysis and data integration of influencer factors in the study area were carried out in geographic information system (GIS) technology, with weights assigned for each factor in combination with a fuzzy analytic hierarchy process, but a low consistency effect was delivered. To optimize the consistency effect, the particle swarm optimization algorithm was introduced, and weights with good consistency effects were obtained. Based on such weights, the suitability evaluation was carried out to select photovoltaic power stations. The evaluation results were compared with those obtained from the fuzzy analytic hierarchy process and the genetic algorithm (GA) optimization model based on FAHP and were further verified in the area of the photovoltaic power stations built. It is demonstrated that the accuracy rate of the suitability of Qujing photovoltaic power station site selection based on the particle swarm optimization model of the fuzzy analytic hierarchy process was the highest at 99.3%.

Existing methods for calculating terrain entropy in grid digital elevation models (DEMs) often face computational anomalies in specific topographies within small windows. To address this issue, an improved method was developed based on the Euclidean distance approach. This method was inspired by Claramunt’s technique of weighting information entropy by the average distance between points with the same value and different values. Specifically, vectors were formed between grid points and categorized by value consistency and relative positions. Those formed between points of different values were classified by the value of the starting point as well as parallel and adjacent relationships. This comprehensive grouping strategy was integrated into distance calculations, becoming a new probability operator that accurately reflects terrain spatial characteristics. Experimental verification confirms that the method proposed aligns with the fundamental concept of entropy, yielding a regression equation of $y=0.011\ln x+0.463$ with a coefficient of determination of 94.73%, a reliability of 44.015, and a measurement ability of 0.757. For the mixed iterative images with gradually increasing spatial disorder, their entropy values should follow a logarithmic trend. Therefore, a logarithmic function is used for fitting. A determination coefficient greater than 50% indicates that the method adheres to the original definition of entropy and is effective in capturing the increasing spatial disorder of the grid DEM. A lower reliability value suggests smoother data computation between the two iterations. A lower measurement ability value indicates slower convergence for grid DEMs with gradually increasing spatial disorder. The improved method was also tested on simulated and real DEMs, and the results showed a strong correlation between calculated terrain entropy values and terrain complexity. By effectively capturing spatial information changes, this approach overcomes the shortcoming of computational anomalies and demonstrates high reliability in terrain entropy calculation in grid DEMs.

Accurate extraction of mountaintop areas from grid digital elevation models (DEMs) is essential for terrain analysis, geomorphological research, hydrological modeling, natural disaster monitoring, and emergency communication site selection. However, existing deep-learning-based methods often suffer from inadequate representation of local details and limited global contextual awareness, leading to blurred boundaries and reduced segmentation accuracy in complex mountainous terrains. To address these limitations, this study proposes a dual-encoder and global-context-enhanced TransUNet framework, named DEGC-TransUNet, for automated mountaintop delineation. The architecture integrates a convolutional encoder to capture fine-grained local terrain features and a MaxViT-based encoder to model multi-scale global context by encoding low-dimensional topographic attributes such as slope and curvature. A dedicated feature fusion module harmonizes complementary representations from both encoding paths, while a BiFormer-based strategy is introduced at the bottleneck to strengthen long-range dependencies and enhance convergence. The experimental results demonstrate that DEGC-TransUNet significantly outperforms baseline models such as TransUNet, DE-TransUNet, and GC-TransUNet, with relative improvements of 19.8% in Intersection over Union (IoU), 10.4% in overall accuracy (ACC), and 10.9% in F1-score. These findings provide a robust solution for mountaintop extraction, with significant potential in analyzing geomorphological evolution, simulating soil erosion, modeling species distribution in “sky island” ecosystems, and optimizing strategic placements for communication base stations and wind energy infrastructures.