Automated Full-Hemisphere Sensor Alignment (AFSA)

We propose to construct a continuous calibration system that is completely software driven. The system will require sufficient computation resources and access to camera feeds from all the cameras to be calibrated. The Phase I will focus primarily on developing the riskiest portion of the proposal to demonstrate feasibility.  The Continuous Calibration is broken into several sub components with inputs being just the camera feeds and outputs being a calibration correction table for each camera. The core of camera calibration is determining which pixels in one camera map to which pixels in another camera. With that knowledge, hand offs from wide field cameras to narrow field cameras and gimbal alignments can be computed. However, the typical process is to either use a calibration box/calibration board, polarized light, or artificial stars in a controlled environment to manually align the pixels between cameras. Similarly, bore-sight alignment is typically done by adjusting the mounting brackets by hand to minimize the offset between the cameras. Here, we propose an automatic, continuous calibration process that can be done en-route and in uncontrolled environments.