Leveraging Computer Vision for Emergency Vehicle Detection-Implementation and Analysis

Abstract

Recent advances in Computer Vision technology have revolutionized the field of Intelligent Transportation Systems. The applications are far reaching- right from traffic monitoring systems to self-driving cars. Most applications entail at least simple, if not advanced image or video analytics at a fundamental level. This paper is an attempt to examine the use of object detection and instance segmentation for emergency vehicle detection, which is indispensable to any Intelligent Transportation System. More particularly, emergency vehicle detection can be programmed into autonomous vehicles as well as traffic signal controllers for preferential signal switching upon encountering emergency vehicles. The architectures implemented are Faster RCNN for object detection and Mask RCNN for instance segmentation. The computational results of these implementations, their accuracies and most importantly, their suitability for emergency vehicle detection in disordered traffic conditions are deliberated. Additionally, the object detection model is contrasted with instance segmentation and the merits and demerits of each are identified, again in the context of emergency vehicle detection.