Artificial Intelligence Applications and Reconfigurable Architectures

9A Relative Study on Object and Lane Detection

Rakshit Jha*, Shruti Sonune, Mohammad Taha Shahid and Santwana Gudadhe

Department of Computer Engineering, Pimpri Chinchwad College of Engineering, Pune, India

Abstract

A self-driving car is a vehicle that can sense its surroundings and navigate without the need for human intervention. It detects environments using a plethora of techniques like radar, LIDAR, GPS, and computer vision. Lane detection is one of the key features of self-driving cars. It is detecting the white/yellow color markings on a surface to ensure that the automobile is within lane constraints. The chapter provides a survey on lane detection approaches, based on “Performance analysis of existing lane detection like CNN based, Hough Transform, Gaussian filter, and canny edge detection and the proposed approaches on different datasets, such as curved roads, big datasets, rainy days, yellow-white strips, day and night lights. The chapter also presents a detailed direct comparison of the You Only Look Once [YOLO] algorithm with Object detection using color masking and provides insight on YOLO algorithms’ predecessors. YOLO is a simple and straightforward algorithm that has a plethora of categories to detect objects live in real-time using a camera, by an input video provided to it and, also in an image given to it as an input. YOLO v3 is a very fast algorithm and was an incremental leap in the domain of object detection. The most noticeable feature in YOLO v3 is ...