House power monitor using OCR(?)

[Jerd]

Hi all!

I would love to get some feedback on this project I'm about to embrace.

I want t use a RPi with a webcam which each hour will take a photo of my house power-meter. Then process the image with some OCR-software (perhaps tesseract would work well?) to extract the power-meter reading. Finaly send the resault to a mysql db.

I plan to use Python to do the local work of fetching information from the OCR software, and do some simple checks to reduce faulty reads as much as possible. Then build a php frontend to show the data - perhaps with jQuery mobile to nicely support phones and pads.

Anyone else doing a similar project? All input is much appreciated.

Best regards, Jon

[tenochtitlanuk]

A quick-and-dirty way is to recognize that a fixed camera position means all seven-segment-bars are at reproducible positions. Just look at a pixel in each bars centre & decode. Much easier than doing proper OCR!

I've an example at http://www.diga.me.uk/sevenSegDecode.html but coded in another language.

[Tafkas]

Interesting idea. Instead of a MySQL db i would recommend rrd-tools. I used a similar approach for recording temperatures in my appartment. The charts are generated on the client side using JavaScript. I also wrote a summary of how I did it.