Edge Detection for Label OCR capability - Currently using OpenCV & Tesseract

edit

I am in the process of building a application which should OCR a rectangular label and need some direction on where/how to start. I am currently using tesseract for OCR functionality and OpenCV for filtering the image for a cleaner read. What I would like to implement is something similar to Card.io

https://github.com/card-io

in which a frame is added and then detect once the label then fills that frame in order to capture the photo. This is an iOS application. Any help would be appreciated. My C++ skills are a little rusty and am looking for the most pain free route to get this done.