Skew Correction, Segment And Compression Of Scanning Image

Nowadays, with the development of computer science and technology and digital image processing technology, many kinds of examinations have used computer online marking system to mark the answer sheet. Compared with traditional marking method, the creation of online marking has made the speed and qualify of the marking more and more excellent. While because of the inaccurate position of the answer sheet or other reasons, the image of answer sheet may lead to skew phenomenon, it can cause many difficulties to image segment and image recognition. In addition, it is vital to the quality of the online marking that whether image segment is accurate or not. Besides, effective image compression is also important to the saving of image of answer sheet. In view of these problems, this paper will be mainly discuss and research image skew correction, image segment and image compression.First of all, the paper introduces several image skew correction algorithms, and interprets those theories. About Hough Transform, the most famous skew correction method nowadays, the paper points out its faulty and limitations in practical application. And based on that algorithm, the paper proposes a improved algorithm, image skew correction algorithm based on Wavelet-Transform and Hough-Transform. This algorithm turns aimed image into binary image first, then we use wavelet transform to get the general picture of the binary image. After we do edge detection and hough transform, then we can get a angle which is the skew angle.Final rotate image. Proved by experiment, this algorithm can reduce the amount of calculation and time of the algorithm effective. So it can make large scale marking more facilitate.In addition, the paper introduces several image segment algorithms too. By so far, it has not produced an uniform image segment standard, so this paper proposes a method that suited to answer sheet image segment according to practical application. This method take advantage of computer auto recognition and artificial mark to segment the image.Besides, the paper also discusses some image compression algorithms, and introduces the current researchs at home and abroad. As regards compression of answer sheet image, this article compares the compression efficiency of the algorithm to explore a more suitable answer sheet image compression algorithm.In the end of the article, the paper summarizes the whole research first, then make a further outlook of the subject.