Scene parameter estimation for the performance analysis of image processing systems

We have implemented scene parameter estimation as an analysis tool for the performance of video-based environment sensing for road vehicles. The dimensions of the scene parameter space are light intensity, weather conditions, and light distribution. In addition to the scene parameters, image errors (blooming, synchronization errors and noise) are recognized. Scene parameter estimation is a key tool also for the development of a a test strategy in view of a safety validation of advanced driver assistance systems.