Processing And Recognition Of The Letters Block Image

Character recognition is one of classic topics which has been studied in-depth. It is one of the most successful applications in recognition. The recognition of letters and numbers is especially important. Because of their small types, it can help to deep study and validate some new theories and methods. Furthermore, the recognition of letters and numbers are extended to some related fields easily, such as motor vehicle license plate recognition, ship and tank number recognition, note paper recognition and so on. With in-depth study, and their use is also more and more widely.The key steps of character recognition are image enhancement and denoising, image positioning and segmentation and character recognition. In this paper, it has a deep and detail study of this three keys. Image enhancement is to enhance the contrast of image, and to improve the quality of image. This study used gray stretch, histogram correction and median filtering to enhance the image. It strengthens the interested part of image and removes or weakens the uninterested part of image. Positioning and segmentation is also the key of character recognition, only when the letters depart from the background it can be identified. In this paper, the image position is used Hough transform. It can scan the image and segmentation letters directly.Character recognition is the key in this study. It studies the theories of character recognition and several existing algorithms. For the characteristics of the letters in this subject, it improved template matching algorithm based traditional algorithms. Achieve a rapid and effective identification. Finally, It studies about camera calibration to determine the positions of the letters and the camera, and complete the work of robots crawling.