• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 29 subscribers
  • Views 628 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DLPLCR4500EVM: Recommended Cameras to begin evaluating the 4500EVM for TILDA-00254 Guide

John Traynor
Intellectual 885 points
Part Number: DLPLCR4500EVM

Hi all,

I am new to this forum so please forgive me if some of these questions are very simple that I am asking

  1. In relation to the guide TILDA-00254 which is a couple of years old now what camera is recommended to be used in order to get up quickly up and going to start evaluating the Lightcraft 4500? I see a reference to this model in the forums - Flea®3 FL3-U3-13S2C-CS 1/3" Color USB 3.0 Camera but this model is only a 1.3 MP. Do I not need something like a 5MP version to get best results from the 4500 chip or is 1.3MP OK?
  2. Should I use a mono or colour camera? What are the pros and cons of using each? (Speed of acquisition and processing in my case is not an issue)
  3. It is difficult to get other cameras working. I have worked with the Genie models from Teledyne Dalsa in the past so it would be great to get some feedback if these models are easy to integrate with this guide
  4. What lens is recommended? Initially I am looking to measuring a part that is about 300mm (W) x 300mm (L) x 50mm (H) in size but will also have a requirement to look at much smaller parts down the line.
  5. If I wanted sub 10uM accuracy / resolution what type of setup / changes would be required? (Feel free to put limits on the size of the part that is to be inspected)
  6. Are there visual studio example projects available in c++ / c#? I was mention that were going to be created but I'm not sure if they are available in the SDK

Thanks in advance

JT

  • 0
    Intellectual 885 points
    Hi all,

    Any help with the questions I have above would be greatly appreciated. I am looking to purchase the EVM kit as soon as possible

    Thanks

    JT
  • 0 in reply to John Traynor
    TI__Mastermind 20045 points
    Hello JT and sorry for the delay,

    1. The indicated Flea camera has been used to achieve reasonable results. You are right that a better camera could obtain better results though. Theoretically a higher resolution camera would get you better results, and depending on the algorithm used, you would want a camera with at least 2x the resolution (so in this case 2*912*1140 pixels). However, you may need to fine tune the algorithm to fully take advantage of this.
    2. There is no advantage to a color camera if you are using a gray coded technique or sinusoidal technique to obtain 3D data. Usually a monochrome camera is cheaper.
    3. I don’t have any experience with that model; however, in general any open CV compatible camera should work fine. You may not have as much functionality as the Flea camera though (this isn’t necessarily due to an actual hardware limitation but is instead due to the fact less code was developed for other cameras. This code could be implemented seperately).
    4. The lens requirement will be to ensure the entire projected image is in the field of view of the camera. Ideally matching the aspect ratio helps but this isn’t required (you will just not receive any information from the camera pixels that don’t have a projected image). To later scan a smaller object you will want to have a larger focal length projected image and camera lens to follow the above requirement.
    5. Accuracy is dependent upon many factors. The key factors are the distance to the object, baseline (distance between the camera and projector), and focal length (projector and camera both matter). The distance to the object squared is inversely proportional to high accuracy. The baseline is directly proportional to high accuracy. And focus is directly proportional to high accuracy. Therefore, with a large baseline, a close object to scan, and larger focus length, you will get higher accuracy. However, there are many other system factors to take into account to guarantee a certain accuracy. I would recommend looking at some research papers that do a good job talking about this.
    6. We do not have any visual studio examples. We just have the ALC SDK.

    Thanks,
    Kyle