Histogram based enhancement of low exposure videos using linear regression and linear programming

Abstract

Nowadays, digital cameras and surveillance systems are more and newlinemore prevalent, whether they are used for information gathering or newlinesurveillance. Researchers are also becoming increasingly interested in newlinebuilding a smart city due to the rapid growth of intelligent video monitoring newlinesystems. The importance of video analytics has grown significantly, newlineparticularly in surveillance applications such as illegal parking warnings, newlinehelmet wear analysis, anomaly detection, stabbing action detection, smoke newlineand fire detection, person re-identification, face detection, and so on. Such newlineapplications may receive their input from a polluted environment or from newlinecapturing devices with a lower configuration. The most prominent problem newlinewith the surveillance system is considered to be the quality degradation of newlineimages and video under varied poor lighting situations. Due to this, object newlineanalysis in such images and videos seems difficult. Detecting, tracking, and newlineenhancing moving foreground objects constitute the major step of any video newlineanalysis process, regardless of the area of use. Consequently, the goal of this newlineresearch is to successfully and efficiently detect, extract, and enhance moving newlineobjects in videos by processing video frames. newlineIn recent years, much research has been done in the field of newlinebackground subtraction and foreground extraction in surveillance video. Also, newlinethere have been many research works in the field of image and video contrast newlineenhancement. Before these strategies could be used in practice, there were newlinestill numerous difficulties and barriers to overcome. Thus, it calls for effective newlineand machine learning-based approaches that can automatically accomplish the newlinetasks of object recognition and enhancement. newline