Smart Document Alignment: Detecting and Correcting Skew in Scanned Images

In document digitization, skew in scanned images—caused by misaligned paper or scanner distortions— can hinder OCR accuracy. This project presents an automated method to detect and correct such skew using Python. The process begins with image preprocessing, including grayscale conversion and edge detection using the Canny algorithm. The Hough Transform is then applied to detect lines and estimate the skew angle based on their orientation. For text-rich documents, Tesseract OCR is also used to estimate skew. The image is then rotated using affine transformation to correct its alignment, improving the accuracy of OCR and subsequent data extraction.