基于高分一号卫星遥感影像的城市绿地提取对比研究
|
摘要
针对目前面向对象方法在高分辨率遥感影像中提取绿地专题信息的特点,文中提出一种利用高分辨率遥感影像提取城市区域范围绿地信息的方法,结合K-T变换和ICA变换,根据地物的遥感影像特征、光谱特征信息和基于阀值的分类技术进行有效波段最优组合及地物分类,从而大幅提高了绿地专题信息提取的精度。研究中提出本方法和基于NDVI的典型绿地提取方法的提取结果进行精度评价,实验结果证明,在城市区域范围尺度上,该方法计算简便且实现了94.97%的高精度和总Kappa系数为0.919 5的评价结果。
For extracting the characteristics of the green project information by the object-oriented method in high resolution remote sensing images,in the paper,a method of extracting green land information of Urban area by using the high resolution remote sensing image was put forward,combining with K-T transform and ICA transform,according to the features of the remote sensing images,spectrum characteristic information and based on classification techniques of the threshold to carry out effective band optimal combination and objects classification,thus the extraction accuracy of green monographic information was improved substantially. Precision of the extracting results of method proposed in the study and the typical green extraction method based on NDVI were evaluated,the experimental results showed that in urban area scale,the calculation method was easy and simple achieving 94. 97% accuracy and the evaluation results of total Kappa coefficient that it was 0. 9195.
For extracting the characteristics of the green project information by the object-oriented method in high resolution remote sensing images,in the paper,a method of extracting green land information of Urban area by using the high resolution remote sensing image was put forward,combining with K-T transform and ICA transform,according to the features of the remote sensing images,spectrum characteristic information and based on classification techniques of the threshold to carry out effective band optimal combination and objects classification,thus the extraction accuracy of green monographic information was improved substantially. Precision of the extracting results of method proposed in the study and the typical green extraction method based on NDVI were evaluated,the experimental results showed that in urban area scale,the calculation method was easy and simple achieving 94. 97% accuracy and the evaluation results of total Kappa coefficient that it was 0. 9195.
引文
[1]许康,卢刚,郭雪瑶,等.基于资源三号卫星遥感影像的城市绿地提取[J].测绘与空间地理信息,2015,38(5):85-88.
[2]李成范,尹京苑,赵俊娟.一种面向对象的遥感影像城市绿地提取方法[J].测绘科学,2011,36(5):112-114,120.
[3]Hame T,Heiler I,Miguel-Ayanz J.An unsupervisedchange detection and recognition system for forestry[J].International Journal of Remote Sensing,1998,19.
[4]苏志成,吕宏伟.基于独立分量分析的遥感影像分类方法[J].科学技术与工程,2007,(23):6244-6247.
[5]郭娜,刘剑秋.TM遥感影像植被信息提取的最佳波段组合选择-以福建省松溪至建瓯高速公路为例[J].福建师范大学学报(自然科学版),2012,28(1):103-107.
[2]李成范,尹京苑,赵俊娟.一种面向对象的遥感影像城市绿地提取方法[J].测绘科学,2011,36(5):112-114,120.
[3]Hame T,Heiler I,Miguel-Ayanz J.An unsupervisedchange detection and recognition system for forestry[J].International Journal of Remote Sensing,1998,19.
[4]苏志成,吕宏伟.基于独立分量分析的遥感影像分类方法[J].科学技术与工程,2007,(23):6244-6247.
[5]郭娜,刘剑秋.TM遥感影像植被信息提取的最佳波段组合选择-以福建省松溪至建瓯高速公路为例[J].福建师范大学学报(自然科学版),2012,28(1):103-107.