Image processing for in-vehicle smart cameras

Smart cameras are increasingly being deployed in many automotive applications. The in-vehicle environment presents unique challenges to the camera functions and can significantly degrade the performance of smart cameras. In this paper, we present two image processing algorithms that address two of the most common challenges for in-vehicle smart cameras, namely exposure control and motion induced distortion. The proposed algorithms are very efficient and highly suitable for embedded real-time smart cameras. The algorithms were first developedfor a real-time in-vehicle automatic license plate recognition smart camera but the techniques are also applicable to many other in-vehicle smart camera applications.