Diffraction images

High-precision displacement measurement plays a significant role in industrial manufacturing. Displacement measurement systems based on optical interference principles typically feature complex optical paths and high costs, with measurement accuracy largely dependent on the precision of the instruments and their stability in varying environments. Based on the sensitivity of optical diffraction patterns to the position of the diffracting object, this paper proposes a novel nano-scale displacement measurement method based on deep learning with diffraction images. The measurement system consists of a light source, diffraction aperture, and image sensor, featuring a simple structure and low cost. This method utilizes a deep learning model to establish a mapping relationship where diffraction images are the input and displacement values are the output. During operation, an image sensor captures the diffraction images and feeds them into the deep learning model to determine the displacement variation of the diffraction plate in the direction perpendicular to the optical axis.