No abstract

Computational imaging adjusts the traditional optics of a camera in order to process the light signal prior to digitalization. A major problem is the effort of subsequent adjustments of the optical components, limiting such cameras to very narrow fields of application. This article analyzes a camera that uses a transparent liquid crystal display instead of a mechanical aperture. The simple adjustability of the optical transfer function enables a flexible application and adaptation to varying ambient conditions. The performance of the camera is exemplariliy analyzed in the scope of depth estimation with the depth-from-defocus method. Our algorithm estimates the scale of the point spread function in a single image by comparing the power spectral density of the observed image with the known aperture shape used to capture the image. The analysis of a test scene shows that the camera is capable to reproduce previous simulative results and that object distances can be reliably estimated from single images.

No abstract

Autonomous driving is a promising technology to, among many aspects, improve road safety. There are however several scenarios that are challenging for autonomous vehicles. One of these are unsignalized junctions. There exist scenarios in which there is no clear regulation as to is allowed to drive first. Instead, communication and cooperation are necessary to solve such scenarios. This is especially challenging when interacting with human drivers. In this work we focus on unsignalized T-intersections. For that scenario we propose a discrete event system (DES) that is able to solve the cooperation with human drivers at a T-intersection with limited visibility and no direct communication. The algorithm is validated in a simulation environment, and the parameters for the algorithm are based on an analysis of typical human behavior at intersections using real-world data.

This project has been funded within the priority program 1835 "Cooperatively Interacting Automobiles" by the German Research Foundation (DFG). Mapping data has been provided by the City of Karlsruhe.