Sapphire growth defect visual detection method based on deep learning

Abstract

The invention discloses a sapphire growth defect visual detection method based on deep learning, and belongs to the field of sapphire preparation detection. In the process of artificial sapphire synthesis, the detection of sapphire growth defects mainly relies on manual detection, which is time-consuming and laborious, the environment in the growth furnace causes difficulties in artificial detection of growth defects, and artificial observations are prone to errors. The invention provides a visual detection method for sapphire growth defects based on deep learning theory, and the method collects no less than 2 million of images of sapphire crystal growth stages by using a camera for processing, generates a training set, sets up a deep learning network, adjusts parameters to train the deeplearning network to generate a training model, uses the training model to detect the images to be tested, determines the crystal growth state in real time, and operates the seed crystal rod to make the growth defect crystal grow normally.

Description

technical field [0001] The invention relates to the field of sapphire preparation and detection, more specifically, it relates to a deep learning-based visual detection method for sapphire growth defects. Background technique [0002] Sapphire is the best material for the manufacture of infrared optical windows and fairings for electronically guided high-speed fighters and missiles. However, natural sapphires are very scarce in nature, and cannot well meet the quality and size requirements of the industry, so most of the sapphires on the market are synthetic. [0003] In the synthetic sapphire process, an important step is to detect the growth defects of sapphire. However, in practice, manual detection of sapphire growth defects is time-consuming and laborious, and the furnace environment of different crystal growth furnaces is different, which makes it difficult to detect growth defects by traditional methods. The traditional method mainly relies on human eyes to observe ...

AI Technical Summary

Problems solved by technology

[0004] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a deep learning-based visual detection method for sapphire growth defects. The invention is based on the theory of deep learning and is mainly used to detect whether there are defects in sapphire during artificial inoculation or synthesis.

Images