I have images of document containing either arabic or english text as image is it possible somehow to know what language out of this two an image contain
Yes, it is possible. You set two possible recognition languages ("English", "Arabic"), then do OCR and check the recognized text character set. Latinic alphabet will indicate English.