Image segmentation is one of the most important steps in the image processing. For example, when image segmentation is used in the field of character recognition, we should separate the text from the image background. However, due to the low image resolution, the complex image background or other reasons, the effect of image segmentation is often not optimal, which also indirectly increases the difficulty of character recognition.
In view of such situation, this thesis introduces the concept of computer vision and details a typical computer vision library OpenCV with its historical development, content, structure and applications. At the same time, this thesis also describes how to improve image segmentation algorithm with the functions from OpenCV library, which aims at correcting the positioning and segmentation results of the printed texts, especially the handwritten texts in a complex image background condition.
At the end of this thesis, we test the effectiveness of the improved image segmentation algorithm with a number of image files and analyze the specific effects of different related texts examples and compute respectively the Improvement rate of the handwritten texts and the printed texts.
Keywords: Image Segmentation, Computer Vision, OpenCV, Text Positioning
目录