摘要
遥感图像空间分辨率的提高,在极大丰富地物目标信息含量的同时,也使得一些传统的目标提取方法受到较大挑战。该文结合基于对象的思想和Boosting算法,提出一种新的针对高分辨率遥感图像中建筑物自动提取的方法。该方法通过构建对象网络关联图像分割和识别,有效解决了一般方法中采用预先定义形状和大小的滑动窗检测目标时效果不佳的问题。然后针对建筑物的目标特性训练有效特征分类器,并利用标记置信度来综合分析图像的各类信息,完成目标提取及后续处理。实验结果表明,该方法可用于提取多种类型和结构的建筑物,准确率高、鲁棒性好,具有较高的应用价值。
Many traditional target extraction methods encountered a new challenge as the spatial resolution is increasing quickly. For the purpose of extracting buildings automatically in that circumstance, a new method combing both the object-based approach and boosting algorithm is proposed in this paper. The method associates segmentation with recognition by constructing a hierarchical object network, which improves effectively the problem of detecting targets with a modifiable sliding window existed in other methods. Then some useful features are selected automatically to train a validate classifier, and the confidence in each label incorporating kinds of information is computed to complete the extraction procedure. Competitive results both for multiform and complicated buildings demonstrate the precision, robustness and effectiveness of the proposed method.
出处
《电子与信息学报》
EI
CSCD
北大核心
2009年第1期177-181,共5页
Journal of Electronics & Information Technology
基金
国家部委基金资助课题
