Inspection and Repair of Optical Damage in Tradition and Deep Learning (Invited)

This paper mainly introduces the inspection and repair methods of optical damage in high power laser system. Because the optical element damage is common in the Final Optical Assembly (FOA) of high power laser system and has great influence on the normal operation of laser system, it is necessary to inspect real time and repair in time, so as to achieve the purpose of recycling optical elements. In online inspection, Final Optical Damage Inspection (FODI) is an important method, which can image and analyze the damage of optical components in real time. In addition, there is an indirect way to obtain damage images, which is to detect the damage by diffraction ring. The size and location of the damage point can be calculated by the relevant formula. For the detection of smaller damage, the tool of deep learning, which can process a large amount of data, is an indispensable method for studying this problem at present. The on-line detection device proposed by them has been a very effective means of detection. With the development of deep learning in image and data processing, convolutional neural networks and decision trees are used to identify and judge the location and size of damage points, so that we can quickly find the damage points. Accurate detection and identification is a premise for the protection and recovery of optical components, then, damage repair needs effective technical means to repair the component and bring it back to the original quality standard as much as possible. The main method of repairing damage is Rapid Ablation Mitigation (RAM), which is the most common and effective method of repairing damage. The premise and key to the damage site treatment is to accurately locate smaller damage points and classify different types of damage so as to determine the subsequent repair steps. Of course, different application scenarios require different technical means. Finally, the structure and flow of optical element damage detection and repair are introduced by the optics recycle loop strategy. This process is very helpful to realize multiple utilization of optical components, save cost and improve utilization rate. Damage detection and repair is an important part of optics recycle loop strategy. Influenced by deep learning in the field of image processing, it is believed that more and more methods of deep learning can be used in researches related to damage detection and repair. In a word, optical component damage detection has developed towards the direction of online detection to improve resolution, and deep learning to help improve classification accuracy and accurate positioning. Damage inspection and repair is an important and indispensable part of the optics recycle loop strategy.