The pixel rectangle index used in object based building extraction from high resolution images

35th International Symposium on Remote Sensing of Environment

地球科学;生态环境科学

Cui, W.H.^1 ; Feng, X.^1 ; Qin, K.^1

School of Remote Sensing and Information Engineering, Wuhan University, 129 LuoYu Road, Wuhan, China^1

关键词: Building extraction;    Extraction accuracy;    High resolution image;    High resolution image segmentation;    Optimal theories;    Principal directions;    Remote sensing images;    Rotation invariant;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/17/1/012233/pdf
DOI  :  10.1088/1755-1315/17/1/012233

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

An improved high resolution object-based building extraction method based on Pixel Rectangle Index is presented in this paper. We use Minimum Span Tree optimal theory to realize object-based high resolution image segmentation. First, we proposed a rotation invariant Pixel Rectangle Index by introducing the principal direction of homogeneous area. Second, we improved the calculation of edge-weight by introducing the band-weight and Pixel Rectangle Index. The QuickBird high resolution images were used to do the building extraction experiment. The experiment result proved that this method can obtain high extraction accuracy and this algorithm can be efficiently used in remote sensing images.