TFT camera acquisition

[josesarra]

Hi everyone.

This is my first time here.

I work for an automobile company and I'm developing automated test for cluster validation (hardware and software).

I'm using Vision Builder for AI 2013.

Everything is running fine when I need to validate telltales. Every led and shapes are being captured by the VBAI.

The main problem that I'm having is related to the TFT (3,5" graysacle) acquisition and OCR validation.

The thing is that every time I make an acquisition, the TFT illumination changes and this changes make the OCR reading intermittent.

Things I have done so far:

• Adjusted the EXPOSURE to the TFT refresh rate
• Removed the NOISE with a LOCAL AVERAGE filter (9 iterations)
• Training the characters only 2x

Is there any step I should take to improve the TFT acquisition?

My setup:

• Camera: Basler acA3800
• Lens: Basler 12mm - 1:18 - C125-1218-5M
• NI Vision Builder for AI 2013
• Vector software to simulate CAN environment

I'd appreciate some help just to stabilize the TTF acquisition.

There is another doubt: how can I adjust the OCR training to avoid the "union" between letters?

Thanks in advance.

[orlandiisabelle]

Hi, Jose Sarra.

What do you mean when you said: "The thing is that every time I make an acquisition, the TFT illumination changes"? Do you know the refresh rate of the TFT? Did you set the exposure period to be greater than the period of a refresh?

About the adjust of OCR training: why are you acquiring the letters united? Low resolution of your camera? Any image processing blurring the image?

I'll wait for your answers.

Best regards,