학술논문
A Machine Vision-Based Person Detection Under Low-Illuminance Conditions Using High Dynamic Range Imagery for Visual Surveillance System
Document Type
Conference
Author
Source
2022 Third International Conference on Smart Technologies in Computing, Electrical and Electronics (ICSTCEE) Smart Technologies in Computing, Electrical and Electronics (ICSTCEE), 2022 Third International Conference on. :1-6 Dec, 2022
Subject
Language
Abstract
The most difficult component of any computer vision application objects recognition and tracking. Video surveillance is a major study subject in computer vision in a dynamic environment, especially for security purposes. Video surveillance technology is critical in preventing crime, terrorism, and other threats. Several existing technologies can accurately monitor a person's mobility in interior environments. As well as regulated outdoor settings, object recognition and tracking at night, on the other hand, remain significant challenges for visual surveillance. The objects are usually farther away, inconspicuous, and have low brightness against the background. In this paper, detecting objects specifically people under low lighting conditions with the help of deep learning algorithms was used. For applications like in-car cameras and surveillance systems that operate in low light, these algorithms might improve picture identification performance.