Derivation of numerical criteria for the occurrence of natural phenomena in video sequences

In this paper are presented the results from an analysis of the information, obtained from dataset of one year video records. The results show that some natural phenomena could be detected by simple numerical criteria. The records have been taken by outdoor surveillance cameras. The applied data processing aims to determine criteria for the occurrence of events affecting the visibility in images. The technique uses dark channel prior images (DCPI) and correlated color temperature (CCT) for limited areas located at identical positions within consecutive frames.