Abstract
Large data bank of images has been accumulated during atmospheric cloudinessground-based observations from 2017 to 2020 near Fryazino city, Moscow Region. The problemof obtained images classification into several types is considered. The types are clear sky (noclouds); cumulus cloudiness of various extent (humilis, mediocris, congestus); very powerfulclouds such as cumulonimbus or nimbostratus; stratus cloud cover; stratocumulus clouds; lightand high-positioned clouds (altocumulus, cirrus, cirrostratus and cirrocumulus). A qualitativeanalysis of the key features of gray, blue, green and red level co-occurrence matrices for variouspixel distances and directions (angles) is performed to solve the problem. An image classification algorithm based on these and other key features retrieved during image preprocessingis developed. The quality evaluation is performed. The developed software tool is currentlysuccessfully used for atmospheric radiometry problems support.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
