Extended Range Detection of Driver Assistance System

Abstract

In this thesis a driving assistance companion system that uses a vision-based detection system to find the detection range of automobiles or mobile robots to offer drivers with important information. Targets include traffic signs, vehicles, animals, and pedestrians, to name a few. Two cameras make up the suggested system. The first camera has a short focal length lens for detecting and tracking pedestrians in the near-mid range, while the second camera has a long focal length lens for detecting and tracking pedestrians in the mid-to-far range. Find detection results of both cameras and eliminate repetitive measurements. After the detection, the tracking procedure is utilized to track and predict the pedestrian's future movements and direction and two algorithms that generate alert and hazard alerts have been created to prevent vehicle-target collisions. This system can be thought of as an advanced driver-assistance system that helps you learn faster and drive better. The helper companion uses detection and monitoring of moving pedestrians from close to far range to predict forthcoming events on the route. Our system improved learnability, driving performance, and reliability, according to test results and mathematical model based on the fundamental physics of the camera and lens is built.