Detection of Hohlraum Target Position for Laser Fusion Experiments

A hohlraum is a cylindrical structure that holds a laser fusion target at the National Ignition Facility. It must be aligned properly for all the 192 laser beams to hit the target and cause a fusion reaction. Video images of the hohlraum are used to align the hohlraum to the required position. A matched filtering based approach is used to locate the circular alignment fiducial of the hohlraum. One of the challenges of the automatic alignment algorithm is the presence of a number of nearly concentric features from which only one will provide the valid position information. The problem is compounded by blurring of relevant features by defocus or insufficient illumination and amplification of non-relevant features. It is shown that to identify the appropriate fiducial; the shape (or size) in addition to amplitude of correlation peak must be considered. Key word: Pattern recognition, image processing, correlation shape, laser alignment