COMPARISON OF THRESHOLDING OTSU AND MORPHOLOGY OPENING PREPROCESSOR FOR IMAGE TEXT DETECTION USING TESSERACT

Abstract

Now everything is completely digital, so image processing is very necessary for this era, one of which is image text processing. processing image text and converting to text using OCR and Tesseract help. There are many helpful applications like this on the internet, but if we want to use them on personal documents it will be very dangerous to use applications from the internet. so it is safer if you use your offline application. This program will not only convert the text converter on the image into text format but will also recognize the text later. So data processing will be easier when the final result is in txt format then for the detection of text in the image will use the help of an open source library from python, namely tesseract. but before using this tesseract library, we have to do preprocessing on our image, because tesseract is very sensitive if there is noise in the image, so here I will use preprocessing thresholding binary otsu method and use dilation. This method will later be compared the final result will be more accurate if detected using tesseract. The final result of this project will detect the text in the image, from here it will detect every word in the image, not every sentence, and will also save the text in the image that has been detected using tesseract