Alignment Of Higher-Order Mode Solid-State Laser Systems With Machine Learning Diagnostic Assistance

Increasing focus is being placed on the development of solid-state laser systems that utilise higher order transverse modes for benefits to applications as varied as material processing [1], particle manipulation [2] and high-resolution imaging [3]. Various computational methodologies have been employed previously to attempt to analyse modal composition, but typically suffer from high computation time and complex experimental arrangements [4,5]. These methodologies can provide an estimate of modal composition; however, they are too slow to provide real time alignment guidance. Development of laser systems operating on these higher-order modes would benefit from the ability to receive feedback from a simple diagnostic system in real-time. Here we report on the development of a convolutional neural network (CNN) to provide laser mode composition estimates using single plane intensity images. Our CNN provides a modal composition estimate in less than 3 ms on consumer grade hardware - a computation time on the order of the upper-state lifetime of many common laser materials. We present the development of the CNN and application examples for alignment of solid-state laser systems operating on higher-order modes.