OCR makes it possible for the user to edit or search the document’s contents. In this paper we describe a special water fill technique for detecting the upside down text document. Each character has a upside and downside filling capacities. A character may have two sides or one side filling capacity or zero filling capacity. The total upside and downside capacities for the scanned page calculated and the page with bigger downside capacity decided to be upright. The merit of the algorithm is that it requires only simple arithmetic operations per image pixel. Our experimental results, based on detecting inversion for 100 documents demonstrate a high detection performance of more than 98%, indicating the validity of the proposed methods.