Questions about combining laser scans with camera photos to form a model?

I wonder if this particular data from laser scans can be combined with camera photos?

It depends what kind of laser scan did you use. It seems like a image with projected texture. Am I right? Also, you need to capture images and laser scan in same time in this case.

HI

Ondrej Trhan CR

Can a laser scan of this data combined with a camera’s photo build a good model? Because the data scanned by the laser is black and white

Sorry, but what kind of laser scan is that? It depends. But theoretically, yes, you can use black and white laser scan data and images to create a model. In most cases, the laser scan data are used for mesh creation, the images for textures. 

But it is quite hard to scan a person in the same position with laser scan and images, as you are moving during laser scanning, which will create a not so good mesh. And also you need to keep the same position as during image capturing. Also, if it is not a terrestrial laser scanner there could be issues using it in RealityCapture, as RealityCapture supports terrestrial laser scanners where it can use the position and also ordered point cloud.

This is infrared laser scanning, I plan to install more than 10 at the same time to scan portraits

use it to scan portraits, so there is no object that will move

 

HI

Ondrej Trhan CR

yeyuchuan 1, do you have some another question?

The pictures I posted are data from infrared laser scans, and I plan to install more than 10 structures for scanning portraits at the same time
Do the modeling so that the model doesn’t look so bad. My question now is can this infrared image data be modeled?

HI

Ondrej Trhan CR

The question asked yesterday, can I use multiple such photos for modeling, can the photos be black and white for modeling? Is it okay to take a set of photos from various angles with the camera for textures?

Hi,

theoretically yes, images like this can be used for model creation and then you can use the colourful images for texturing. But these images looks like they are not good quality or resolution… Also for me the model looks unfocused. Maybe it will be better idea to use data projector as I mentioned before. What is the resolution of these images?

https://www.youtube.com/watch?v=dKX48fLuyv0&list=PL56jeA0rCS3Lf1OI1a_Z7GGqSgTLEsk8V&index=16

Do you have information on this company? I see they don’t seem to use projection。

Hi, all I know that it is The Scan Truck. It is possible that they don’t use projector, but they have calibrated cameras and I think they also know the position of each camera (XMP workflow) which helps with alignment and don’t need to use texture projection for image alignment. Also, in this rig is more than 200 cameras, which is also helpful.

 

HI

Ondrej Trhan CR

Projector + camera to take pictures, is there any good suggestion for this interval?
Because what I use to control the camera is xangleCS.
After taking the photos with projection, you need to wait for the transfer to complete before taking the second group.

The best way is do it right after, as you should keep the pose for model creation.

Like here: https://www.youtube.com/watch?v=HK5-DkYYhpA, 1:30, it is captured almost in one moment.