Surface reflectance data obtained by the absolute atmospheric correction of satellite images are useful for land use applications. For Landsat and Sentinel-2 images, many radiometric processing methods exist, and the images are supported by most types of commercial and open-source software. However, multispectral KOMPSAT-3A images with a resolution of 2.2 m are currently lacking tools or open-source resources for obtaining top-of-canopy (TOC) reflectance data. In this study, an atmospheric correction module for KOMPSAT-3A images was newly implemented into the optical calibration algorithm in the Orfeo Toolbox (OTB), with a sensor model and spectral response data for KOMPSAT-3A. Using this module, named OTB extension for KOMPSAT-3A, experiments on the normalized difference vegetation index (NDVI) were conducted based on TOC reflectance data with or without aerosol properties from AERONET. The NDVI results for these atmospherically corrected data were compared with those from the dark object subtraction (DOS) scheme, a relative atmospheric correction method. The NDVI results obtained using TOC reflectance with or without the AERONET data were considerably different from the results obtained from the DOS scheme and the Landsat-8 surface reflectance of the Google Earth Engine (GEE). It was found that the utilization of the aerosol parameter of the AERONET data affects the NDVI results for KOMPSAT-3A images. The TOC reflectance of high-resolution satellite imagery ensures further precise analysis and the detailed interpretation of urban forestry or complex vegetation features.