what features should have an image to be used with OCR

[Xontrol]

Hello everyone. I would like to know what features should have an image to be used with OCR. I have managed to implement the OCR (with the VDM module) but applying a filter that turns me Threshold the image (bmp) to another in original black and white, just so I can work. When I try from labview not allow me to work with the original BMP file. Note the images are captured from a webcam.

Thank you.

[muks]

If you have a colored image (RGB) convert it to a greyscale image. Use color plane extraction and try extracting luminance plane.

[Xontrol]

Thanks Musk. I have a question, what should be the treatment he should give a color image to convert TIFF finally, I mention this because the examples in Labview / VDM / OCR use files in this format.
Thank you very much.

[muks]

No there is nothing like you have to tiff only. You can very well use a Jpg image as well.