摘要
基于ETM+遥感影像,运用ERDAS IMAGINE 9.2和ENVI4.7对上海市绿地信息进行了提取。实验过程中首先对图像进行预处理,然后尝试了分别用初始波段组合法、主成分分析法(PCA)、土壤调节指数法(SAVI)和波段指数综合法对上海市绿地信息进行了提取。根据研究区域的实际情况,得出植被景观最好的是土壤调节指数法,然后利用以上4种方法所得的假彩色合成图像进行监督分类,用目视判读的方法对ETM+影像的分类结果进行精度检验,得出波段指数综合法分类结果最好,该方法是一种能有效减弱土地背景影响的绿地信息提取方法。
Based on ETM＋ remote sensing image,the information of urban green space of shanghai was extracted by ERDAS IMAGINE 9.2 and ENVI 4.7.In the process of the experiment,image preprocessing was accomplished at the first step.Four methods for green space extraction were compared.They are original bands combination method,principal component analysis method（PCA）,soil adjusted vegetation index（SAVI） method and the combination of original bands and derivative bands method.Comparing the four images created by the four methods,the soil adjusted vegetation index method created the best visual effect according to the actual situation of Shanghai.Supervised classification was carried out with four false color composite images created by four methods,and accurate assessment was done by ERDAS IMAGINE 9.2.The results show the combination of original bands and derivative bands method has the best accuracy in four methods,and it is an effective method to reduce soil background effect for the extraction of green space.
出处
《电子测量技术》
2010年第12期4-7,23,共5页
Electronic Measurement Technology
基金
上海市重点学科和科委重点实验室项目(S30108
08DZ2231100)
上海市科委科技攻关计划(10511501702)
上海市科委国际合作项目(08590700500)
国家自然科学基金(61071185)
