Image Processing Technologies In Laser Damage Threshold Testing

Driver is the key component in controllable inertial confinement fusion. As an Inertial confinement fusion driver , the laser always works at extremely high power, Usually, it exceeds kilo-joule. To some extent, the power level and the performance of the laser system are limited by the damage resistance of the components to intense laser illumination. Therefore, it is vital for the researchers and technicians working in high power laser area to get the precisely damage threshold of optical components. However, the laser-induced-damage in the optics is so complicated, for it is affects by lots of factors iying in pahysics or chemistry. It has been attracting the researchers' attention.This thesis focus on image processing technologies' applied to the laser damage threshold testing. Firstly, we review traditional laser damage detection methods. nevertheless, some of them are off-line detection methods. The author brings forward three different detection algorithms based on image processing. In experiments, we manage some irridated optial thin film Images by author'methods and get satisfying results, these results are in accord with ISO11254 and GB/T16601.Secondly, considering the inaccurateness of optical spot measurement by blade method, the author brings forward optical spot area cauculation method based on improving 2D Hough transform. In experiments, we retreat optical spot image and calculate its' area by this method. We get satisfying results.The next, we analyse the thin film defects' sorts, characters and cause of formation and bring image segment algorithm based on inheritance 2D maximum Entropy. We segment thin film defect image successfully by this method and statistics defect density by mathematics morphologic calculation for binary image. We can infer the defect sorts according to defect density, this is very important to improve thin film qualities.Finally, we introduce dynamic image collection system, image manipulation Software development and bring forward a scheme about damage inline detection.