Neural Network-Based Peri-Urban Zone Delineation and Resilience-Oriented Ecological Security Membrane Planning: A Case Study of Zhengzhou, China

Peri-urban zones are critical interfaces where urban expansion, agricultural production, and ecological processes overlap. However, ecological-network planning in these areas is often constrained by uncertain boundary definition and insufficient integration between habitat quality and landscape connectivity. Taking Zhengzhou, China, as a case study, this paper proposes a resilience-oriented Ecological Security Membrane planning framework that links peri-urban boundary delineation with the prioritization of ecological sources, corridors, and critical points. A deep neural network was used to distinguish the urban core, urban fringe, and peri-urban zone from multi-source land-use and socioeconomic indicators, achieving an overall classification accuracy of 93.1%. Priority ecological sources were then identified by coupling biodiversity quality, patch morphology, area thresholds, and connectivity contribution, while corridors and critical points were prioritized to support network reinforcement. The results reveal a peri-urban ecological structure characterized by source concentration in the western mountainous and eastern agroforestry areas, insufficient ecological continuity along the Yellow River corridor, and key bottlenecks at transport and urban-expansion interfaces. The proposed framework advances peri-urban ecological planning by translating source–corridor–node analysis into a spatially explicit planning structure. Future research should test the robustness of this framework under multi-year, multi-seasonal, and scenario-based urban-growth conditions.