The Study Of Automatic Selection Of Parameters In OCR Image Capture

Current OCR (Optical Characters Recognition) systems have overcome the disadvantages of time-consuming and hard-sledding with the high error rate when the information was input into computer manually. The characters and images were inputted into computer automatically with high rate. The implementation of automatic input was based on the clear original manuscript. Therefore, it is important that suitable parameters are selected on the image scanning stage.This paper carries on the research on the automatic parameter selection of the scanner on the stage of image capture through the analysis of the pertinent parameters and TWAIN protocol carefully. The pre-scanned images are analyzed by the method of gray histogram. Parameters are adjusted properly according to the features of histogram. In order to discover the resolution that are suitable for scanning on the basis of determine the character size by the projection and over come the question caused by unsuitable parameters artificially. On the basis of be loyal to the original manuscripts and be good to recognition that the clear images of scanned documents are obtained. In view of the comparative analysis of different documents, the experiment indicates the quality of the original manuscripts is guarantee according to the adjustment of the parameters automatically.