摘要
快速获取城市植被的胁迫状态,不仅对城市植被健康状况的维护,而且对城市生态环境的改善具有重要意义.在对受胁迫植被的生理特征和光谱特征进行分析的基础上,利用星载高光谱Hyperion数据,计算出与胁迫相关的14种高光谱植被指数,在此基础上运用BP神经网络算法建立了城市植被胁迫强度分类器,对城市植被的胁迫强度进行了识别与分析.结果表明:城市中心商住区的植被受胁迫程度明显高于城乡结合部和郊区;植被的受胁迫现象在大块绿地外围呈环状分布;构建的植被胁迫强度分类器能够较为准确地反映植被受胁迫的强度信息,可为大面积城市植被胁迫监测提供一种较为可靠而快捷的方法.
To quickly obtain the information of urban vegetation stressed level is of great significance in maintaining urban vegetation health and improving urban eco-environment. Based on the analysis of stressed vegetations physiological and spectral characters, and by using Hyperion hyperspectral data, 14 hyperspectral vegetation indices related to stress were calculated, and a classifier of urban vegetation stressed level was developed based on this calculation and BP Neural network. The application of this classifier in identifying the vegetation stressed level in a case study area of Guangzhou City showed that the vegetations in commercial and residential districts were apparently experienced higher stress than those in suburban regions, and the stressed level showed a ringy distribution around large pieces of greenbehs. This classifier was able to quickly and accurately identify the vegetation stressed level, and thus, could be used as an effective tool in monitoring urban vegetation stressed condition.
出处
《应用生态学报》
CAS
CSCD
北大核心
2007年第6期1286-1292,共7页
Chinese Journal of Applied Ecology
基金
国家杰出青年科学基金项目(40525002)
中国博士后基金项目(20060390208)
"985工程"GIS与遥感的地学应用科技创新平台重大项目(105203200400006)
国家自然科学基金项目(40601010)
广东省环境科学与技术公共实验室开放基金资助项目(060207)