OCR for pictues of a Bathroom scale

Requires. Python2 , SimpleCv --

This is a prototype for software to read common machine information using a smartphone camera. The jpg from the phone has accurate date and time and so all we need to do is upload pictures to a server and parse the picture for the numbers we are interested in.

This project also serves to demonstrate instrumentation and regression testing. The testloop.py script reads a txt file which contains the expected output and the file-name to process. It verifys that the algorithm is returning the correct numbers for each picture and keeps track of elapsed time currently around 5 sec/per picture.

the weight.py script analyses a single picture.

All of the scripts can be run independently using intermediate files that are output from a prior stage. Running testloop produces a TRB file which contains a list of all images that failed the test. TRB is in a format suitable for reinput to testloop.

I have not included my daily script which uses the scale object in weight to update statistics that I keep in R.