Improving the accuracy of remote sensing image classification by the combination of pixel-based and object-based image classification methods
PDF (Tiếng Việt) | Download: 481

Keywords

phân loại dựa vào đối tượng
phương pháp cây phân loại
kết hợp phân loại dựa trên đối tượng và điểm ảnh Object-based image analysis
random forest
combination approach

Working Languages

How to Cite

Nong, T. O., Tran, X. T., Vu, T. T., & Ta, H. T. (2022). Improving the accuracy of remote sensing image classification by the combination of pixel-based and object-based image classification methods . Journal of Geodesy and Cartography, (53), 48–56. https://doi.org/10.54491/jgac.2022.53.655

Abstract

Remote sensing images are increasingly important in providing information to monitor the earth's surface. To make the most of this data source, remote sensing image classification methods are becoming smarter and more efficient, converting information on images into valuable information. The object-based classification method has proven its accuracy compared to the pixel-by-pixel classification method. However, this method is often applied to high-resolution satellite images but has not been widely applied to medium-resolution satellite images, moreover, the combined use of both classification methods has not been proven. lots of testing. This study combines object-based classification and pixel-based classification methods to improve classification accuracy. The results show that combining both methods can improve the overall classification accuracy by 20% when using high-resolution images.

https://doi.org/10.54491/jgac.2022.53.655
PDF (Tiếng Việt) | Download: 481

Downloads

Download data is not yet available.