基于SOM神经网络的秦岭北麓保护区域村庄分类与发展策略  

Village classification and development strategy in the north foot of Qinling Mountains based on SOM neural network

作　　者：赵哲 吕楠 姜翠梅 ZHAO Zhe;LYU Nan;JIANG Cuimei(College of Urban and Environmental Sciences,Northwest University,Xi'an 710127,China;Xi'an City Planning and Design Institute,Xi'an 710082,China)

机构地区：[1]西北大学城市与环境学院,西安710127 [2]西安市城市规划设计研究院,西安710082

出　　处：《桂林理工大学学报》2023年第4期608-616,共9页Journal of Guilin University of Technology

基　　金：陕西省重点领域科技创新团队项目(2020TD-029)。

摘　　要：保护秦岭生态环境是推进生态文明、维护生态安全、促进秦岭范围内人与自然和谐共生的重大举措。协调秦岭北麓区域内生态保护与镇村发展间的矛盾,是未来秦岭生态环境保护工作的重中之重。为了合理制定秦岭生态保护与区域内村镇建设规划方案,构建考虑生态保护区特点及乡村自身属性特征村庄的多维度乡村分类评价指标体系,引入SOM神经网络方法建立了适用于秦岭北麓区域生态特点的乡村分类模型与指标体系,基于区域内475个村庄的调查数据识别乡村类型,提出了针对不同乡村特征的发展策略。研究表明:经过充分训练的SOM神经网络分类模型可以有效识别不同村庄的相同特征值属性,村庄类型识别精度较高,该模型具有良好的适用性和科学性,为乡村政策制定和优化提供了支撑。Protecting the ecological environment of Qinling Mountains is significant to promote ecological civili-zation,maintain ecological security and promote the harmonious coexistence between man and nature.Coordina-ting the conflict between ecological protection and town-village development in the north foot of Qinling Moun-tain is of the top priority in Qinling mountain ecological environment protection in the future.The environmental protection of Qinling Mountains and the construction of towns and villages are planned.A multi-dimensional ru-ral classification and evaluation index system are constructed based on the characteristics of ecological reserves and rural attributes.SOM neural network method is established with index system suitable for the regional envi-ronment.Based on the survey data of 475 villages in the region,the rural types are identified,and the develop-ment strategies for different rural characteristics are implemented.The research shows that the SOM neural net-work classification model can effectively identify the same eigenvalue attributes of other villages,with good rec-ognition accuracy.The model provides support for rural policy formulation and optimization.

关 键 词：秦岭北麓区域 乡村分类 评价指标体系 SOM神经网络算法 发展策略 

分 类 号：TU982.29[建筑科学—城市规划与设计]