Image correction method, device and system

Abstract

The invention discloses an image correction method, device and system. The method comprises the following steps: performing gray processing and adaptive threshold binarization processing on an image to obtain a binarized image; extracting contours in the binarized image, and determining the contour forming the maximum communication area as an image outer contour in the contours; carrying out straight line detection on the outer contour of the image, identifying side lines of the outer contour of the image according to pre-acquired image shape features and detected straight line features, and determining intersection points of the side lines as corner points of the image; and performing image correction according to the angular point position of the image. According to the invention, the binarization processing method of the adaptive threshold can ensure that different thresholds are adopted in different brightness areas of the image with the reflection condition, so as to overcome theproblem that the image information cannot be collected due to the reflection condition of the image.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image correction method, device and system. Background technique [0002] Optical character recognition technology (OCR character recognition technology) is mainly a technology that translates the shape of characters in a recognized image into computer text. This technology is widely used in various scenarios that require automatic acquisition of text information in images. In order to recognize characters accurately, before using OCR character recognition technology, it is necessary to correct the image to eliminate the irrelevant background area, image skew and other problems in the image, which requires first identifying the edges and corners of the target object in the image points to distinguish the target object from the background in the image. At present, there are three commonly used corner detection methods: corner detection based on grayscale image...

AI Technical Summary

Problems solved by technology

The above-mentioned corner point detection method can accurately identify the corner points in the correction process of general images, but in practical applications, due to problems such as shooting angles and photo materials, images often have reflective areas. In the process, it was found that for images with reflective areas, the above-mentioned corner point detection method often has the problem of multiple recognition and false recognition, which leads to poor image correction effect and causes troubles for subsequent character recognition

Images