Ai enabled blind spot detection using rcnn based image processing

Abstract

Due to the rapid increase in the rate of road accidents and traffic density, modern automobiles are equipped with intelligent systems like Adaptive cruise control and Lane Departure Warning System. Therear-view mirror can be effective to observe a limited range only and there are zones that cannot be viewed. This region is referred to as the blind spot. Therefore, we present a method to detect the vehicles from the side and the rear for Blind Spot Detection with vision system incorporating RCNN. Blind spot detection is a key technology among driver aids that provides 360 degrees of electronic coverage around the car during motion.The methodology presented in this paper uses two stereo cameras as input devices which constantly capture the images at the blind spot area and the information is passed to the main controlling unit. Potholes are also detected and the alert is sent to the nearby vehicle. The incorporation of Artificial Intelligence would help in enhancing the picture quality and blur or cancel the background images probable of misreading the target image. RCNN is used for the vehicle detection and for evaluating the relative distance between the vehicles.This technology allows us to provide a realistic environment for commercial vehicle drivers as they can’t monitor the side and rear-view mirrors all the time, making the whole driving experience more comfortable.