基于移动窗口法与 ANN 模型的澄迈林场森林景观研究.pdfVIP

第35 卷 第12 期 中 南 林 业 科 技 大 学 学 报 Vol. 35 No. 12 2015 年 12 月 Journal of Central South University of Forestry Technology Dec. 2015 Doi:10.14067/j .cnki.1673-923x.2015.12.017 http : // 基于移动窗口法与ANN模型的澄迈林场 森林景观研究 陈冬洋，陈彩虹，常双双 （中南林业科技大学，湖南长沙4 10004 ） 摘 要：利用ARCGIS 和Fragstats 软件，基于移动窗口法和ANN 模型，对澄迈林场森林景观进行研究分析模拟。 结果表明：适宜的景观分析粒度为60 m 栅格，最佳样方尺度为540 m；人工神经网络模型适合森林景观模拟研究， 收敛效果理想。将ANN 模型分析与移动窗口法结合，精确研究分析森林景观格局，为林场景观格局量化分析提 供参考，为林场经营管理提供科学合理指导。 关键词：森林景观模拟；景观指数；移动窗口法；粒度效应；ANN 模型； 中图分类号：S718.5 文献标志码：A 文章编号：1673-923X(2015) 12-0095-07 Forest landscape study for Chengmai Forest Farm, Hainan province based on moving window method and ANN model CHEN Dong-yang, CHEN Cai-hong, CHANG Shuang-shuang (Central South University of Forestry and Technology, Changsha 410004, Hunan, China) Abstract: Using ARCGIS and Fragstats software, based on the moving window method and ANN model, the forest landscape of Chengmai Forest Farm in Hainan province was simulated. The results show that the appropriate landscape analysis granularity is 60 meters (landscape grid), the optimal quadrat scale (shrub) is 540 meters; artificial neural network model have the ideal convergence effect and is suitable for forest landscape simulation. The combination of ANN model and moving window method, can accurately analyzed forest landscape pattern, thus providing a reference for quantitative analysis of forest landscape pattern and more scientific and rational guidance for management of forest farm. Key words: forest landscape simulation; landscape inde