Two-stage document image nonlinear distortion correction method based on projection method

Abstract

The invention discloses a two-stage document image non-linear distortion correction method based on a projection method. 1, binarizing the obtained document image, adaptively detecting and extractingthe boundary of an effective region by using a boundary detection and extraction algorithm, and obtaining a document image correction region; 2. For the preprocessed document image, the baseline of the text line being extracted based on the projection method, and then the base point is selected to construct the distorted surface transformation model for the first stage correction, which is calledrough correction. 3, after that coarse correction is carried out, the baseline of the text line being extracted from the image by the projection method again, and then the connected regions in the text lines being correct in a more detailed second stage according to the extracted baseline of the text lines in a text line unit of behavior, and the fine correction is called the fine correction. Theinvention can overcome the influence of a series of non-linear distortions such as perspective, tilt, distortion and the like in the document image on the OCR identification of the document image, andimproves the accuracy of the OCR identification.

Description

technical field [0001] The invention belongs to the application field of image processing and correction, and specifically proposes a two-stage document image nonlinear distortion correction method based on a projection method. Background technique [0002] With the continuous development of digital technology in modern life, more and more information is transmitted and preserved through digital information. However, traditional information carriers still mainly rely on paper media. When converting paper media information into digital information, artificial The way of entry is time-consuming and laborious. The development of OCR technology solves the problem of character recognition in the conversion of paper media information into digital information, thus greatly improving the efficiency of converting paper media information into digital information, but it is mainly used in scanners, which are bulky and slow on traditional instruments. However, the more common smart de...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects in the above-mentioned document image nonlinear distortion correction, the present invention proposes a two-stage document image nonlinear distortion correction method based on the projection method, which makes full use of the local and global features in the document image, staged and efficient Improve the correction effect of document images taken by mobile devices, thereby improving the accuracy of OCR recognition

Images