Moving Object Detection in Night Time: A Survey
Anu Singha, Mrinal Kanti Bhowmik · 2019
The object monitoring performance is greatly depends on the enhanced quality of images. The quality of these images is affected at night video systems due to low contrast of salient objects as well as several atmospheric conditions such as fog, rain, dust etc. As a consequence, the intensity, color, polarization, coherence of scenes alters. The awful appearance of night images under subjective lighting and atmospheric conditions is a general problem for analysis in computer vision. In this paper, we have aimed to provide a survey work to researchers under four categories: (i) review on change detection datasets either thermal or visual-thermal, (ii) review on feature extraction based object detection techniques, (iii) review on background modeling based object detection techniques, (iv) review on convolutional neural network (CNN) based object detection techniques. This paper will benefits for those who require a details review over change detection datasets and detection techniques for moving object detection related applications.