A method and device for image registration of industrial printed matter based on deep learning

Abstract

The invention discloses a method and device for image registration of industrial prints based on deep learning. Specifically, according to corresponding cutting rules, slices of the same size are respectively extracted from a template image and an image to be registered to obtain a group of slices. Image slice pairs are obtained after deep fusion, and the image slice pairs are input into the registration network model for training to obtain the registered image slices; the edge parts of the registered image slices are discarded, and then the slices are aligned and spliced ​​to obtain a complete Registered print image of . The framework of the registration network model of the present invention is a UNet-like network, and the spatial transformation layer is used to perform non-rigid registration on feature maps of different scales, and the registration feature map is fused with the feature maps of adjacent scales in the decoder, and at the same time, the adjacent scales are aligned The quasi-field is also fused, which comprehensively improves the registration ability of the model for images of larger deformed prints. The invention can solve the problems existing in image registration of some industrial paper printed matter at present, and the registration effect is good.

Description

technical field [0001] The invention belongs to the field of defect detection of industrial printed matter, and in particular relates to a method and device for image registration of industrial printed matter based on deep learning. Background technique [0002] Image registration is an important research direction in the field of computational vision. The main purpose is to match two or more images under different conditions (illuminance, shooting angle and shooting position, etc.) to achieve geometric alignment. At present, image registration can be divided into traditional registration methods and image registration methods based on deep learning. Traditional registration methods include traditional rigid registration methods and traditional non-rigid registration algorithms. Among them, traditional rigid registration algorithms, such as feature-based image registration algorithms, search grayscale, edge and artificially designed image parameters in the image parameter sp...

AI Technical Summary

Problems solved by technology

However, paper printed matter is a non-rigid material, and it is easy to produce various deformations under the action of external force. Even if it can be fixed by clamps, it is difficult to obtain the exact same shape as the template image. Traditional registration methods are difficult to effectively solve paper printed matter. The problem of non-rigid image registration cannot fundamentally reduce the misjudgment caused by defect detection

Images