基于国产GF-1遥感影像的面向对象桥梁提取方法研究

以国产“高分一号”2m/8m高空间分辨率遥感图像为数据源,使用基于规则的面向对象的方法实现了对高分辨率影像中桥梁目标的精确提取。首先,经过多尺度分割实验并结合下垫面特征选择最优分割尺度；其次,利用水体指数、阈值函数等方法建立规则集,逐步获取水体和桥梁潜在区的矢量文件；最后,通过二值化、数学形态学处理、叠加分析等方法成功提取桥梁目标。实验结果表明,该方法可以准确、高效地提取出桥梁信息,总体精度在90%以上,Kappa系数在85%以上,在高分辨率遥感影像高精度提取桥梁的实践中具有广泛的适用性,该研究成果或对国产高分影像处理系统的研究与应用提供了一定的科学参考。

Object-oriented Extracting Bridges Information Based on China-made GF-1

Based on China-made GF-1,high spatial resolution remote sensing image as data sources,using object-o-riented method based on rules to achieve the goal of high spatial resolution image in the bridge of the extraction. First,through a large number of multi-scale segmentation experiments,the optimal segmentation scale according to the underlying surface feature selection;Secondly,useing of water body index,the threshold function method to es-tablish rules,such as gradually access to water and the potential bridges vector files;Then,by using binarization pro-cessing,mathematical morphology algorithm,overlay analysis and so on operation,The ultimate success of extraction bridge targets. Results that the method can accurately and efficiently extract the information of bridges,the overall accuracy of over 90%,the Kappa coefficient of more than 85%,the results of the study or to the research and appli-cation of domestic high image processing system to provide some scientific reference.