基于支持向量机的六安市城区多源景观格局演化分析
Evolution of Landscape Pattern Analysis Based on Support Vector Machine of Multi-sources' Satellite Images in Lu'an Urban Area
李陶 1孙贤斌 2胡莎莎 1杨佳2
作者信息
- 1. 安徽省金寨县城乡规划局,安徽六安237300
- 2. 皖西学院环境与旅游学院,安徽六安237010
- 折叠
摘要
通过收集六安市城区多源多时相卫星影像数据,采用支持向量机分类方法和GIS技术,分析了2009-2017年的景观格局演化情况.结果表明:①Sentinel-2影像地物解译特征显著,图像分辨率满足地物类别识别,总体分类精度达92.41％;②在景观结构动态变化中,水域基本保持稳定,建设用地面积增长了约3.4倍,且向东扩展趋势明显;耕地和林地面积总量逐年下降,在东部区域趋于破碎化;未利用地则受建设用地扩大速度的影响,一直处于波动变化中;③在景观格局演化中,水域保持动态平衡状态,建设用地和耕地增加量分别达80.61％和55.64％,林地和未利用地面积减少明显,分别达63.39％和72.33％;④演化后期的建设用地和林地演化速度明显加快,而耕地面积速度开始下降,城市扩展显著.
Abstract
By multi-sources' and multi-temporal satellite images data gathering,the evolution of landscape pattern in 2009-2017 years in Lu'an city was analyzed by using support vector machine classification and GIS technology.The results showed:①The feature interpreting of image object is significant on Sentinel-2 satellite image,also the image resolution to meet the object classes identified standard and the overall classification precision reaches 92.41％.② In the dynamic change of Landscape structure,the water area remained stable,construction area increased by about 3.4 times and the eastward expansion tendency is obvious;the total amount of arable area and forest area decreased year by year in the eastern region and tends to be fragmented;unused areas is affected by the expansion of construction area,which has been in the wave changes.③In landscape evolution,water area maintains the dynamic equilibrium state,construction area and arable area increased respectively by 80.61％ and 55.64％,forest area and unused land reduced significantly by 63.39％ and 72.33 ％ respectively.④In the later stage of evolution,the construction area and arable area has been speeded up obviously,and the evolution rate of arable land began to decline,also the urban expansion is remarkable.
关键词
景观格局/支持向量机/多源遥感影像/六安市Key words
Landscape structure/Support vector machine/Multi-sources'remote sensing images/Lu'an City引用本文复制引用
基金项目
皖西学院2016年度校级科研项目(WXZR201635)
皖西学院2016年度校级质量工程项目(2016wxxy31)
皖西学院2016年度校级质量工程项目(2016wxxy10)
出版年
2017
国家科技期刊平台
NSTL
维普
万方数据