High Spatial Resolution Remote Sensing Image Classification Based on Pixel Shape Index Method

• Published in: Journal of physics. Conference series 2021-06, Vol.1952 (2), p.22054
• Main Authors: Kong, Hui, Liu, Jin Bao
• Format: Article
• Language: English
• Subjects: Feature extraction; High resolution; High spatial resolution; Image classification; Image resolution; Physics; Pixel shape index; Pixels; PSI; Remote sensing; Spatial resolution; Spectra
• ISSN: 1742-6588; 1742-6596
• DOI: 10.1088/1742-6596/1952/2/022054

Then the pixel shape index method is used to extract the pixel shape index features of high spatial resolution remote sensing images and supplement the spectral features.The experimental results show that the pixel shape index feature can effectively distinguish the ground objects with similar spectral features but different geometric shapes, and is superior to the spectral feature classification method in accuracy.Compared with the small ripple feature method and the multi-scale region feature method, the pixel shape index method also achieves better results.On the other hand, it is found in the experiment that the method is easily affected by the detailed information in the high spatial resolution remote sensing image, and the classification effect of the region rich in detailed information is not ideal.In this paper, pixel shape index method is used to effectively distinguish ground object targets with similar spectral features but different shapes, and extract pixel shape index features of high-resolution remote sensing images. Compared with spectral features, pixel shape index features have great advantages in accuracy, and it is also conducive to the supplement of spectral features.