Ask Your Question
0

Detect 7 Segment fault on a LCD display

asked 2015-12-13 20:10:04 -0600

renanfonteles gravatar image

Hello there,

I have to develop an algorithm to detect faults on a LCD display with 7 segment characters. I started to develop the algorithm , but all I did til now was finding the Region of Interest (ROI) , the display.

This is a screenshot of the algorithm execution: image description

Thera are 4 displays in the picture, and I find the regions where they are positioned. The display will be at the state with all the segments turned on, so it should be easy to detect any fault on them.

Now I have to finally execute an algorithm to find faults on the segments, but I'm not sure how am I suppose to do it. I was thinking in using SVM and machine learning for detecting the characters, but I wonder if this is the best way of doing this, as I don't need an OCR for reading each character, I just have to know if all the segments are on or not.

Would you guys have any advice in how should I make this application?

Thank you very much.

Renan Fonteles.

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
4

answered 2015-12-14 09:49:16 -0600

pklab gravatar image

You should be able to do much better with alignment and binarization. After this, use a pre-set mask to detect all 7 segment and use logical OR to detect fault.

digit image description OR mask mask = result result

digit image description OR mask mask = result result

Black pixel in the result reveals faults.

edit flag offensive delete link more

Comments

Hello there,

Thank you very much for replying. I get it, that's exactly what I need, an optimal and simple solution . I will try this right now. Thank you again :D

renanfonteles gravatar imagerenanfonteles ( 2015-12-14 10:41:16 -0600 )edit

nice to see. btw very good alignment and digit segmentation is a must for this solution.

pklab gravatar imagepklab ( 2015-12-14 11:00:18 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2015-12-13 20:08:46 -0600

Seen: 2,599 times

Last updated: Dec 14 '15