S&M2720 Research Paper of Special Issue
Published: November 17, 2021
Development and Comparative Analysis of Geospatial Feature Automatic Extraction System in Open-source Environment [PDF]
Dong Gook Lee, Ji Ho You, and Hyun Jik Lee
(Received February 24, 2021; Accepted May 14, 2021)
Keywords: open source, geospatial feature extraction system, development, classification accuracy
The aim of this study is to develop a system for geospatial feature extraction from images to be obtained from CAS500-1/2 satellites currently being developed by the Ministry of Land, Infrastructure, and Transport, Republic of Korea. The feasibility of automatic geospatial feature extraction is verified by applying the relative radiometric normalization technique to KOMPSAT-3A satellite images, which are expected to have similar specifications to CAS500 images. Furthermore, the developed system is compared with commercial software to verify its classification accuracy. In this study, two KOMPSAT-3A satellite images were collected and relative radiometric normalization was performed. Identical parameters and threshold values were applied to both the commercial software and the developed system to extract geospatial features by feature class and analyze the classification accuracy using an error matrix. Image segmentation and image classification were performed for grassland, ground, roads, buildings, and water bodies. The results indicated a classification accuracy of 90% or higher, which was set as the accuracy goal. The difference in the classification accuracy of the two systems was less than approximately 1%, implying comparable performances for the two systems. Using the geospatial feature extraction system developed by us, it is expected that basic data will be generated for monitoring national territory such as forests and urban areas.Corresponding author: Ji Ho You, Hyun Jik Lee
This work is licensed under a Creative Commons Attribution 4.0 International License.
Cite this article
Dong Gook Lee, Ji Ho You, and Hyun Jik Lee, Development and Comparative Analysis of Geospatial Feature Automatic Extraction System in Open-source Environment, Sens. Mater., Vol. 33, No. 11, 2021, p. 3729-3744.