The rapid distribution of digital cameras has caused several new problems related to text recognition. Based on experimental studies, was revealed that existing OCR systems cannot cope with complex perspective and geometric distortions that arise when photographing text document. Therefore it is necessary to apply text documents pre-processing so that the text lines were straight and horizontal. This article briefly consider existing methods pre-processing documents and found that it depend on the type of distortion and not universal. Proposed new method involving the mathematical raising of straightened text lines on the image and heterogeneous distortion correction based on a page surface transformation model. This method is better than others because it is universal and corrects any type of distortion, including a combination of several types of distortion.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.