Books To Be Scanned Automatically Correct Image Distortion

The scanner may scan each kind of documents conveniently and rapidly. But when scans thicker books witch were not disconnect, the obtained image can have various distortion, causes person to recognize the writing to be difficult.The result of correction was very good when the parameters which describe the image distortion were known. However, limited to the conditions and measurement tools, it was difficult to measure the distortion parameters precisely. How to obtain more precise distortion parameters becomes a key issue.The paper has studied the characteristic of scanned book image thoroughly, and established the relationship model between distortion parameters of books scanned image and gray variation on the basis of the original elliptic cylinder and plane correction model, and expressed the relationship with a multinomial. The distortion parameters can be calculated automatically according to the gray curve of the margins and the polynomial coefficientsIn view of the questions that existed in the experiment, proposed the method of improving the precision for distortion parameters calculated. Established a gray curve data file which allow the nearest bending state of any scanned book images can be found, higher precision distortion parameters can be calculated with the appropriate polynomial coefficients.A method to correct the scanned book images on different resolution has been put forward. According to the resolution of the scanned book image, carry on sampling or interpolating to the gray curve values which in the gray curve data file before comparing with any gray curves. Enable scanned book images on different resolution all to obtain a better correction.The paper presented the method that calculate the distortion parameters automatically, Programmed and achieved the goal that corrected the distortion of scanned book images automatically. The experiment indicated that, this research has obtained the satisfying correction effect.