Optimal location method of emergency communication nodes based on digital terrain features

Following natural disasters that disrupt communication networks in affected regions, establishing an emergency communication network is crucial for effective rescue operations. There are still challenges in choosing the right location for deploying communication nodes based on real terrain features to ensure regional connectivity. This paper proposes a ‘terrain feature extraction + node optimization site selection’ model based on high-precision digital elevation model (DEM) data. Processing DEM data to highlight terrain features to construct the base data. Then the Precision Contour-Mask RCNN algorithm (PC-Mask RCNN) is used to extract mountain contours, and terrain feature node methods are employed to construct a set of terrain feature points (e.g. peaks, ridges, and contours). By combining the advantages of reinforcement learning and multi-objective optimization algorithms, the PPO-NSGAⅡ collaborative optimization algorithm is proposed using line-of-sight simulated signals to optimize high-quality site selection schemes. Experimental results demonstrate that the model and method can accurately summarize terrain features (mIoU > 0.9) and rapidly optimize high-quality site selection schemes with superior coverage and accessibility than other algorithms (effective coverage [Formula: see text] 0.78, accessibility [Formula: see text] 0.5). When applied to real-world scenarios, the algorithm can provide node locations, coverage status, accessibility, and recommended paths for deployment solutions in emergency rescue situations.