An Intelligent Lane, Object Detection and Speed Estimation for Ego-Vehicles using Perception Algorithms

Autonomous vehicles promise to revolutionize transportation by expanding mobility and improving traffic safety. All of this future auto tech hinges on autonomous vehicle perception - the processing of visual input to vital Advanced Driving Assistance functions such as lane detection, vehicle and pedestrian recognition, and traffic sign detection. The goal of this work is to increase the level of autonomy in ego vehicles, which can drive through challenging traffic situations (completely autonomously). This research will then take on one of the portions with a low latency and high accuracy target, which is safety + efficiency (using ADAS – Advanced Driver Assistance Systems). The system proposes an ego vehicle-based computer vision solution for pedestrian, vehicle, and traffic sign detection along with lane detection. The entire system consists of many small systems like the YOLO-NAS Module, Gaussian Blur Module, Hough line transform, and Canny Edge Detection. A detailed flow diagram is given at the end and is curated during the experiment analysis meeting.