Micro-nano structure electron microscope scanning picture automatic detection method, system and medium

Abstract

The application discloses a kind of micro-nano structure electron microscope scanning picture automatic detection method and system, belong to micro-nano structure detection technical field.The method identifies the geometric center and contour of micro-nano structure unit from original scanning electron microscope picture by visual neural network and image processing algorithm, obtains scale information by OCR algorithm, converts the structure information of micro-nano structure unit into physical unit, carries out flaw identification according to preset information, generates binary mask matrix, then constructs ideal grid dot matrix according to preset period and calculates affine transformation parameter, carries out global feature matching based on simulation transformation parameter, determines the position of original scanning electron microscope picture in design layout, finally carries out error statistics based on the result of matching.The scheme realizes the automatic processing of large quantities of images, accurate positioning and geometric parameter extraction of nanostructure, and supports multidimensional statistical analysis, to provide objective, efficient and traceable technical support for quality control of micro-nano manufacturing.

Images