Fast automatic classification of input and exit surface laser-induced damage in large-aperture final optics

Under the condition of inhomogeneous total internal reflection illumination, a fast automatic classification method based on machine learning is proposed to solve the problem of determining the location of surface damage sites. First, the far-field light intensity distributions of damage sites on the input and exit surfaces are calculated by using numerical calculations. After analyzing the optical properties, we present a feature vector for characterizing all damage sites in the image captured by the final optics damage inspection system. Finally, an autoencoder-based extreme learning machine is used to identify the surface on which a damage site resides. The experimental results show that the maximum testing accuracy of this method is 97.66%. Compared to the method used at Lawrence Livermore National Laboratory, the method proposed in this paper shows improved accuracy and faster speed of classification in practical applications. In particular, the proposed algorithm for damage classification outperforms several state-of-the-art methods on our experimental dataset.