Image correction method and device and server

Abstract

The invention provides an image correction method and device and a server, and the method comprises the steps: carrying out the first image correction processing of a target image based on the image correction method, achieving the coarse-grained correction of linear distortion, and obtaining an intermediate image after the first correction; and then determining an average deflection angle of text characters in the target image, and performing second image correction processing on the intermediate image according to the average deflection angle of the text characters by utilizing the characteristic of regular format of the target certificate in the target image, thereby realizing fine-grained correction for linear distortion, and finally obtaining the corrected target image. Therefore, the method can effectively aim at a target certificate which contains fewer text characters but has a regular format and a target image with linear distortion, is better suitable for various different types of linear distortion scenes, accurately eliminates the linear distortion in the target image, and obtains a corrected target image with better effect and higher precision.

Description

technical field [0001] This specification belongs to the technical field of artificial intelligence, and in particular relates to an image correction method, device and server. Background technique [0002] In some business handling scenarios, customers often need to provide and display documents such as ID cards that contain relatively few text characters to the business handling personnel. [0003] Usually, business personnel will first take an image containing the above-mentioned certificate, and then perform OCR recognition (optical character recognition) on the image to extract and use the relevant text information in the certificate. [0004] However, the quality of the image containing the certificate directly captured by the business personnel is relatively poor. For example, affected by factors such as shooting methods, shooting equipment, and shooting environment, images captured for subsequent OCR recognition often have linear distortion, which affects the accura...

AI Technical Summary

Problems solved by technology

However, the above method requires that the outline of the document in the image must be complete during the specific implementation

In many linear distortion scenarios, the edge of the document outline may also be lost, making it difficult to accurately and effectively implement correction processing for linear distortion based on existing image correction methods

[0006] In addition, for documents containing relatively few text characters, such as ID cards, relatively few features can be extracted and utilized, which further increases the difficulty of correction processing for such images

[0007] For the above problems, no effective solution has been proposed

Images