Design And Implementation Of Image Region Character Segmentation System

Character segmentation is both the foundation of pattern recognition technology and an important research topic in digital information management science.Based on the previous image character segmentation algorithm, the regional character segmentation system of image is designed. The system is used Microsoft Visual Studio2010development platform and C Sharp language to design, and the main contents of the system includes file operation module, correct the relevant content of original image module, image preprocessing module, edge detection module, and character segmentation module. The system satisfy accuracy, real-time, robustness and so on in the design, it is also a services system that based on digital image processing, and better used in pattern recognition, computer vision and other related technology.The tilt when the photographer taking pictures that leads to the character information of the image in a certain degree of tilt. In the system of correction module, it aims to change the content of the tilt to straight. The design of the correction module is to use the quadratic curve equation, and according to the user’s choice of the four basic points to determine the content which within the four basis points is to be corrected. The corrected image is clearer on the vision, and it is also more conducive to later image both at character segmentation and character recognition technology research than the original image.The part of image preprocessing, firstly, it makes the correction image to a gray image; secondly, it makes the gray image to deal with de-noising processing and image sharpening processing. This part highlights the characters information in the image, in addition to the background noise in the image, and finally get clearer and less noisy image.The part of the edge detection module, which is used the relevant classical algorithm for characters in the image to take edge detection. This paper analyzes and compares the results of image which is processed by Sobel operator edge detection, Prewitt operator edge detection and Robert operator edge detection.The part of the image character segmentation module, different segmentation algorithms were lead to different character segmentation results, and general image segmentation algorithm will lead to weak anti-interference ability, poor adaptability shortcomings when to segment the image characters. Based on the connected components analysis and vertical projection theory, character segmentation system accomplished to segment the characters of the corrected image.