Electricity Meter Status detection based on Optical Character Recognition and Image Processing There is two external libraries used in this Project : "tess-two" for OCR and "OpenCV" for the Image Proecessing.
There is two methods implemented in this application the first one is :
-Using a custom Camera with a drawn box on the canvas that covers the surface of the meter.
-Place the box over the meter and take a picture to crop automatically the wanted part and detect the numbers
-Result is not accurate. The quality of the croped image needs to be optimised.
The second method is :
-Take a picture with the default camera.
-Crop manually the picture to start the Image Processing and the OCR automatically in the background.
-Result is accurate.
Screenshots -- Test Object used is a kitchen scale which have the same display as an electricity meter :
