A comparison of the cloud detection results between the UDTCDA mask and MOD35 cloud products

Abstract

UDTCDA (universal dynamic threshold cloud detection algorithm) is a new cloud detection method which was proposed recently. This cloud detection method is supported by priori surface reflectance obtained from MODIS surface reflectance product (MOD09). Reflectance of four bands in the wavelength of visible to near infrared are used to detect the cloudy pixels. Because there is a priori reference data on the ground surface, pixels of thin cloud and the fractional cloud can be well detected from the clear pixels. MOD35 is the MODIS cloud mask product, combined use of reflectance and brightness temperature to determine the cloud pixels at 250m, and 1km resolutions. 22 out of 36 bands in the visible, near-infrared, and thermal infrared bands are used to create a high quality cloud mask. The methods of visual interpretation and comparison with the CALIPSO data are used to evaluate the two cloud detection algorithms.