A COMPARISON BETWEEN 3D HIGH-DEFINITION MAPS CREATED BY PHOTOGRAMMETRY AND BY LASER SCANNING APPLIED FOR AN AUTONOMOUS VEHICLE

Abstract

Seft-driving cars are a key innovation in the automotive industry with many benefits that can be achieved to reduce major current traffic problems such as accidents, jams, parking lots, and so on. Various researchers and companies, especially in developed countries, try to solve many problems such as developing the drive-by-wire system, making mathematical algorithms, applying artificial intelligence, with the hope of bringing autonomous vehicles to life. In order to step by step capture the technology and get ready for the test of autonomous vehicles, 3D high-resolution maps, as an important part of the vehicle’s localization and path planning need to be studied in depth. In this paper, based on the selected mathematical algorithm, the performing of two methods for building 3D high-resolution maps will be analyzed to find out the advantages and disadvantages of each one. The results show that the high- resolution map constructed by using lidar is more accurate and detailed, whereas the map constructed by using images with coordinates is more intuitive. Therefore, to be able to develop autonomous vehicles with high accuracy for the whole city, a mapping method using lidar- camera fusion is essential in which map of the detailed roads is created by Lidar and map of the rest areas is built by optical imaging method.