Photometric alignment for surround view camera system

Surround view camera system is an emerging automotive ADAS (Advanced Driver Assistance System) technology that helps drivers park safer by allowing him/her to see the complete 360-degree surrounding of the vehicle. Such a system normally consists of four to six wide-angle cameras mounted around the vehicle, each facing one direction. From these camera inputs, a composite bird-eye view of the vehicle is synthesized and shown to the driver in real time during parking. One common problem in existing solutions is that the composite view has noticeable boundaries between adjacent views due to different lighting conditions faced by different cameras, as well as different auto-exposure (AE) and auto white balance (AWB) each camera has. This paper describes a computationally efficient method to correct for the photometric misalignment, i.e., brightness and color mismatch, across different cameras. With the proposed methods and proper calibration of the cameras, we achieved a seamlessly stitched 360 degree surround view.