2022/02/05 by Sreenivasa Hikkal Venugopala, Venugopala, Sreenivasa Hikkal
Computer Science · Engineering · #Advanced Neural Network Applications #Artificial Intelligence (cs.AI) #Autonomous Vehicle Technology and Safety #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Industrial Vision Systems and Defect Detection
paper · pdf · doi:10.48550/arxiv.2202.02521
openalex publication_date 2022/02/05 · openalex created_date 2022/05/05 · openalex updated_date 2026/07/28
Estimating and understanding the surroundings of the vehicle precisely forms\nthe basic and crucial step for the autonomous vehicle. The perception system\nplays a significant role in providing an accurate interpretation of a vehicle's\nenvironment in real-time. Generally, the perception system involves various\nsubsystems such as localization, obstacle (static and dynamic) detection, and\navoidance, mapping systems, and others. For perceiving the environment, these\nvehicles will be equipped with various exteroceptive (both passive and active)\nsensors in particular cameras, Radars, LiDARs, and others. These systems are\nequipped with deep learning techniques that transform the huge amount of data\nfrom the sensors into semantic information on which the object detection and\nlocalization tasks are performed. For numerous driving tasks, to provide\naccurate results, the location and depth information of a particular object is\nnecessary. 3D object detection methods, by utilizing the additional pose data\nfrom the sensors such as LiDARs, stereo cameras, provides information on the\nsize and location of the object. Based on recent research, 3D object detection\nframeworks performing object detection and localization on LiDAR data and\nsensor fusion techniques show significant improvement in their performance. In\nthis work, a comparative study of the effect of using LiDAR data for object\ndetection frameworks and the performance improvement seen by using sensor\nfusion techniques are performed. Along with discussing various state-of-the-art\nmethods in both the cases, performing experimental analysis, and providing\nfuture research directions.\n