This paper discusses issues of correction of printed and handwritten texts using mathematical morphological methods, word and letter skew correction, segmentation, localization and filtering of gradient backgrounds of images resulting from uneven illumination. Through the use of developed algorithms, we are able to significantly improve the quality of the original image