Automatic Detection of Low Light Images in a Video Sequence Shot under Different Light Conditions

摘要

Nocturnal insects have the ability to neurally sum visual signals in space and time to be able to see under very low light conditions. This ability shown by nocturnal insects has inspired many researchers to develop a night vision algorithm, that is capable of significantly improving the quality and reliability of digital images captured under very low light conditions. This algorithm however when applied to day time images rather degrades their quality. It is therefore not suitable to apply the night vision algorithms equally to an image stream with different light conditions. This paper introduces a quick method of automatically determining when to apply the nocturnal vision algorithm by analysing the cumulative intensity histogram of each image in the stream. The effectiveness of this method is demonstrated with relevant experiments in a good and acceptable way.