Hi,
I tried high detail reconstruction and i can’t really see benefits from normal detail. So i’m trying to increase max features per image and preselector features to get better result in my mesh (texture is perfect).
Is there a rule of thum here. I tried increasing Max features per image to 80000 and preselector to 30000.
Now i went from 1.3M points to 1.5M and from 3.7M to 4.5M projections. Does that mean that my reconstruction will be more accurate ? What do i need to do to increase the quality of my mesh (get rid of wobbling edge and holes) ?
Thanks a lot.
Hi Henri
This is a useless solution. increasing points for alignment has absolutely minimal impact on meshing results, we work differently from other solutions regarding feature points vs meshing quality…
As for the meshing “issues”, you need well-captured data to get crisp meshing results…
I understand CR works differently.
Can you explain what is the interest to play with max features points and preselector then please ?
Thanks.
Hi Henri
MAX FEATURES define how much points are detected in an image
PRESELECTOR takes the defined amount of points from MAX FEATURES, selections are based on points “quality” and it’s best when you use 1/2- 1/4 from the MAX FEATURES
Thanks
Ok i understand, but if it doesnt make my mesh reconstruction better what does it do ?
henri wrote:
Ok i understand, but if it doesnt make my mesh reconstruction better what does it do ?
Try checking Assessing Alignment Quality in the Help section available in the RC.