3D Orthomosaic From Unmanned Aerial Vehicle (UAV) Image Data For Real-World Environment 3D Map Creation Autonomous Vehicles Simulation System
Muhammad Yudhi Rezaldi, Andri Fachrur Rozie, Dikdik Krisnandi, Al Hafiz Akbar Maulana Siagian · 2021
Testing is necessary to prove the reliability of autonomous vehicles (AV) statistically and to reduce fatal accidents. However, the manual testing process takes a long time and be expensive. One of the solutions to address this issue is to use a simulation system. Thus, it is necessary to create a virtual environment that includes roads, pedestrians, and buildings. The purpose of this research is to find an appropriate and affordable way to create 3D maps and data drive simulators based on the real-world environment. This research is conducted using an experimental method to find out how the 3D orthomosaic effect of unmanned aerial vehicle (UAV) from a drone that has been further processed through the RoadRunner and CARLA/Unreal engine simulation software (Independent Variable) to produce a simulation map for AV (dependent variable). Results of this research prove that the 3D orthomosaic image data generated by the UAV, which was processed through the RoadRunner and CARLA/Unreal engine simulation software, could be used properly in making 3D maps of the real-world environment to become the data drive of the AV simulator.