Log in or register to post comments

Object Target Best Practices?

December 8, 2016 - 4:09pm #1

So I've read through the Object Recognition docs and I just had a few questions. I've 3D printed a model, tried scanning it in various setups, lightings, tried to mimic the perfect lighting scenarios as described, but tracking is very poor. I know that Vuforia's tracking is top notch so I must be doing something wrong. 



1. My model is about 2 inches x 2 inches x 2 inches. Is this too small? Need it be larger?

2. Because I've 3D printed my model, it's all one color. The docs say "The surface of the object should have contrast-based features.' Does this mean I should paint the figure for best results? Or will this not improve tracking?

 

Thank you so much for your time! I appreciate it

Object Target Best Practices?

December 12, 2016 - 2:29pm #2

Hello,

I don't think that your model is too small, but if it is all the same color then my guess is that not enough keypoints were extracted during the scanning process. I agree with your assessment about adding contrast-based features to the object's surface.

To confirm that additional keypoints were added, you can compare the number extracted during scanning from the app.

Hope this helps.

Thanks

Log in or register to post comments