Real-time Image Processing and Restoration Prototype for the SEM: Initial Steps

Abstract

In previous work, we have reported on a series of projects utilizing GPGPUs (General Purpose Graphics Processing Units) to improve instrument performance, notably preliminary work in scanning electron microscope (SEM) autofocusing [1], autofocusing and astigmatism correction using Fourier domain algorithms [2], specific SEM-related image processing techniques [3], and automated gun alignment for tungsten filament emission sources [4]. GPGPU computation is now an established means of exploiting the tens to thousands of processor cores available on modern graphics cards, with the proprietary Compute Unified Device Architecture introduced by NVIDIA in 2006 enabling developers to utilize GPUs by means of extensions to C and similar programming languages (see [5]), and OpenCL (Open Compute Language) providing an alternative open-source framework for executing parallel programs across heterogeneous hardware.