tesseract | [WIP]

Actually, I have to say that tesseract is very touchy to play with. According to my experiences, I can easily say that if you -as a human- are not able to read a text clearly, you shouldn't expect tesseract to read it either.

First of all; to get better results, it is a must to make a good preprocessing. I strongly recommend anyone dealing with tesseract to check their documentation about Improving the quality.

In your case, problem is about the resolution. Is low resolution a reason for tesseract not to read a text ? Answer is absolutely yes. Documentation says:

Tesseract works best on images which have a DPI of at least 300 dpi, so it may be beneficial to resize images.

In here DPI means dots per inch and its suggested lower limit is 300 DPI which is higher than your image. When you resize the image to a higher resolution, for example 10 times bigger:

Now even if DPI satisfies, now you are losing the accuracy and getting noises.

Note: It also doesn't mean that higher resolution means better results. Please check here.

Note: If you really need to continue on these types of images, you may need to have a look at here. First you get higher resolution and then deblurring operation, this may help to figure it out.

So here is how I fixed it

It turned out that system didnt display the correct error message because it couldnt use the library System.Drawing.Common which is not supported by Linux.

Fixed that by using libgdiplux the Linux implementation of System.Drawing.Common

Sometimes the first error message may be misleading without a clear cause

In this case the primary message "The system cannot find the specified file"

Will lead a user to concentrate on why a filename is not correct, as in this case.

What the error should report is that a required file in the dependencies was not found. which can be caused by one or more Tesseract or related Leptonica / Language data files not in the correct location either due to no install or poor install.

It transpired that installing tesseract on windows from https://github.com/UB-Mannheim/tesseract/wiki "the script now works fine"

Note a missing dependency was the cause of a similar message here Import ocrmypdf in Visual Stdio Code in Python

Pytesseract, on its own, doesn't handle table detection very well - the table format isn't retained in the output, which can make it difficult to parse, as seen in your output.

So splitting the table into distinct columns, performing OCR on each, and then rejoining the columns will help. This is slower, but it is more accurate.

Dilation can help, which adds white pixels to existing white areas (using the threshold and image you currently have). This expands the narrow areas of the numbers.

In my experience, to improve the accuracy generally means splitting the table up into different sections, as well as testing different thresholds and dilation settings.

I've found a decent workaround. First off I've made the image larger. More area for tesseract to work with helped it a lot. Second, to get rid of non-digit outputs, I've used the following config on the image to string function:

So I managed to fix the problem with the following line:

Just be sure you've installed the underlying library the Python module is taking advantage of, for example:

You'll have better OCR results if you improve the quality of the image you are giving Tesseract.

While tesseract version 3.05 (and older) handle inverted image (dark background and light text) without problem, for 4.x version use dark text on light background.

Convert from BGR to HLS to later remove background colors from the numbers in the top half of the image. Then, create a "blue" mask with cv2.inRange and replace anything that's not "blue" with the color white.

You can use %sh in a separate cell to execute the shell commands on the driver node. To install tesseract, you can do: