At present, the most basic method for detecting particles on the surface of optical components is visual observation. This is based on the experience of workers to determine the number and size of particles. The final result is often influenced by many subjective factors. Now, there are also some professional inspection instruments, such as: laser shear speckle interferometry to detect small defects on the sheet; projection measurement uses THz spectrometer to track THz pulse delay, and to get quantitative data on the surface condition of the component; C-CT That is, atomic force microscopy, also known as scanning electron microscopy, detects particulate contamination on the surface of optical components. Although the accuracy of the results measured by these instruments is high, it is not suitable for the on-line detection of particles having a size of several hundred micrometers.

At this stage, the researchers designed two systems that could meet the requirements of the NIF damage detection system of the United States: The laser optical component damage detection system (LODI) was used to detect the transmission optical components and the target room vacuum window damage; The terminal optical element damage detection system (FODI) is used to detect the damage of the terminal optical element. Both systems use dark field imaging techniques to increase the contrast between damaged images. The FODI system can detect cracks over 50 microns in size with a 99% confidence level and it takes approximately 2 hours to detect 192 beams. However, the above-mentioned systems are complicated to operate, costly and difficult to carry, and are difficult to apply in practice.

The device designed for detecting large-aperture optical instruments' surface particles in this paper is based on a dark field imaging system and can obtain high-resolution images of LAOEs at far working distances. Adjust the lens magnification with a stepper motor and then auto focus. The device is easy to install, easy to carry, and can be detected online. Camera internal parameters and lens distortion parameters are corrected off-line. Online correction of homography between the current image and the ideal image. The current deformed image can be corrected by a homography matrix and an image correction algorithm. With adaptive threshold algorithm and SSE2 algorithm, binary images with background and extracted particles can be obtained. The particle distribution can be obtained by calculating the number of particles in the sub-regions.

Experimental results show that the compact device can meet the detection requirements of large-aperture optical components, and the process is highly efficient and accurate.

(Original Title: Automated Measurement of Surface Microparticles on Large-aperture Optical Elements)