Image resolution needed for OCR return is a tricky subject. The general rule is a preferred resolution of 300 dpi (dots per inch). However, this generally applies to images of text obtained from a scanner and dpi values can be misleading when a camera is used to image the text. DPI is only relevant if the document size is 1:1 and focal distance will vary according to camera placement. Typically, one inch within the image obtained from a camera is not proportional to one inch within the text document. Font size is another factor that is independent of resolution and can lead to poor OCR results. For instance, a 16pt font at 200 dpi will return better OCR results than an 8pt font at 300 dpi. A better measure of image resolution for OCR purposes is obtained by counting the x-height in pixels of the text. X-height is the height of a lower case x within the document. According to Tesseract (open source OCR program), an x-height of 20 pixels or better is preferred.
