TRUE SENSOR CAPTURED RADIANCE AND SURFACE REFLECTANCE FROM LANDSAT-8 MULTISPECTRAL IMAGE SETS

Abstract

Proper validation of multispectral image sets that are produced by Satellite Sensors, in this era of big-data computing, is an important aspect in signal processing of natural semantic scenes. In this research article we explore fundamental issues that are cropped up in radiance correction and surface reflectance measurements. We calibrate the raw Landsat-8 imagery digital numbers by using standard functional oriented modular programs in R software, for true sensor radiance and reflectance values. This important preprocessing step is essential for activities that use multispectral images for natural resources monitoring and change-detection studies. It is noted that the procedures for preprocessing can be applied in machine learning algorithms for scientific studies.