Colloquium "Computer vision methods for the on-road nighttime vehicle detection"

Abstract:

Driving at night is dangerous as it is demonstrated by research which reveals that approximately 91 percent of nighttime rural fatal crashes occur on dark roadways and 9 percent occur on roads lighted by streetlights. More than a half number of all accidents in rural areas happens during the nighttime. This is usually attributed to the lower performance of the human visual system under insufficient light and depth information. In order to reduce the number of accidents automatic automotive nighttime systems can be introduced which help the driver by providing additional information about current traffic conditions. Those systems are commonly based on headlight and taillight detection which are very often the only cues available during nighttime driving. The most common tasks include: automatic headlight dimming in order to prevent drivers of other vehicles from being subjected to excessive glare levels, time to collision estimation where the system warns the driver about a possible collision and can perform measures to prevent it (e.g. breaking), and relative and absolute speed estimation of other vehicles. Although such systems are used for different purposes their structure differs only in high-level stages. Low-level steps such as image preprocessing, light blob detection, object tracking and even camera calibration are shared. The talk will give an overview of so far developed nighttime driver-assistance systems and methods which are commonly used for the nighttime in-vehicle light detection. The review of this area will serve as a basis for further research.

CV:

Darko Jurić received his BSc and MSc degree from the Faculty of Electrical Engineering and Computing in Zagreb. He finished his undergraduate study program in Computing in 2010 and in 2012 his master study program in Computer Science with a specialization in Image processing and analysis. Currently, he is employed at the Faculty of Electrical Engineering and Computing, University of Zagreb, as a research engineer at the Department of Electronic Systems and Information Processing within Image Processing Group, where he is pursuing his PhD degree. His research interests are in the areas of image processing and machine learning which is used for solving problems in computer vision.