An Indoor Surveillances System Using Fisheye Camera and YOLOV4 Object Detection

Abstract

Currently, most indoor surveillance systems employ conventional cameras in a monitored place to have full coverage. However, there are still uncovered areas under surveillance because of the limitation of conventional cameras, for example, the shooting angle restriction. To alleviate the problem, this paper presents a surveillance system based on a fisheye or 360-degree panoramic camera and YOLOv4 object detection. The proposed approach consists of four stages: (i) capture panoramic images, (ii) convert the captured image into a conventional image, (iii) detect objects, that is, human objects, and (iv) record them if necessary. The proposed approach will reduce costs, save installation time, and eliminate areas uncovered in a surveillance system.