Actions
Forum Thread Details
  • State Suggested Answer
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 392 subscribers
  • Views 961 views
  • Users 0 members are here
Related

Which real time image processing environment is best?

farooq123

I am doing my FYP based on machine vision. That will identify the faulty products running on the conveyer belt and they will get removed. I need help in following things.

 

* Which real-time image processing IDE will be suitable for my project?

* which algorithms will be best? (right now I'm testing some samples on Matlab, just for understanding purpose)

* which microcontroller will be best when I'm switching towards embedded system?

 

Your help will be appreciated. Thanks.

Parents
  • 0

    You need to provide us with a lot more information about your requirements before we can provide you with guidance.

     

    Image collection and processing can be very easy or very difficult depending upon you application.

     

    So please take some time to elaborate on what you want to do.

     

    Then we can help you find the hardware and software that will enable you to proceed.

     

    DAB

  • 0 in reply to DAB

    Thanks for responding to my question.

     

     

    My requirements are:

     

    1) Which algorithm is best for detecting a faulty product based on colour and shape w.r.t predefined image?

     

    2) Which microcontroller or processor will be suitable for this project?

     

    3) Which digital camera should I use for real-time image capturing purpose?

     

    4) My project should be fast, as a conveyer belt moves with somehow fast speed so the faulty product should be removed within 2 seconds, maximum.

  • 0 in reply to farooq123

    A best solution is subjective.

     

    To try out your ideas, I would use a PC or RPi with a webcam and use OpenCV software.

     

    That basic configuration would let you capture images in various pixel resolutions so that you can see how pixel resolution affects detection time.

     

    Depending upon your conveyer speed, you will find you need to balance pixel size for detection probability.  Simple color should be easy, but the more complex the shape, the more processing time you will need for reliable data.

     

    DAB

  • 0 in reply to DAB

    As we know that products coming on conveyer belt have different orientation but we have a single predefined image with the fixed orientation, so is there any algorithm which is orientation invariant?

Reply Children