Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLI-B7, 579-583, 2016
https://doi.org/10.5194/isprs-archives-XLI-B7-579-2016
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.
 
21 Jun 2016
COMPARISON OF PIXEL-BASED AND OBJECT-ORIENTED LAND COVER CHANGE DETECTION METHODS
Zhenlei Xie1, Ruoming Shi1, Ling Zhu1, Shu Peng2, and Xu Chen1 1School of Geomatics and Urban Information, Beijing University of Civil Engineering and Architecture, Beijing, China
2National Geomatics Centre of China, Beijing, China
Keywords: Land Cover, Change Detection, Pixel-based, Segmentation scale, Object-oriented, 30m resolution Abstract. Change detection method is an efficient way in the aim of land cover product updating on the basis of the existing products, and at the same time saving lots of cost and time. Considering the object-oriented change detection method for 30m resolution Landsat image, analysis of effect of different segmentation scales on the method of the object-oriented is firstly carried out. On the other hand, for analysing the effectiveness and availability of pixel-based change method, the two indices which complement each other are the differenced Normalized Difference Vegetation Index (dNDVI), the Change Vector (CV) were used. To demonstrate the performance of pixel-based and object-oriented, accuracy assessment of these two change detection results will be conducted by four indicators which include overall accuracy, omission error, commission error and Kappa coefficient.
Conference paper (PDF, 1646 KB)


Citation: Xie, Z., Shi, R., Zhu, L., Peng, S., and Chen, X.: COMPARISON OF PIXEL-BASED AND OBJECT-ORIENTED LAND COVER CHANGE DETECTION METHODS, Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLI-B7, 579-583, https://doi.org/10.5194/isprs-archives-XLI-B7-579-2016, 2016.

BibTeX EndNote Reference Manager XML