Photogrammetry Test 1 - First attempt at photogrammetry

As part of the testing for the 703 unit, I decided to create a photogrammetry scan, this starts to enable me to learn how to use the software as well as collect the right quality images. For the test, I used a dice tower diorama that I had made, I set this up on a turntable with soft lighting so as not to create shadows. 

I used a Canon camera on manual and used a lens set at 50mm zoom so as not to get any distortion. 

After this I took lots of images of the tower and the base, I made sure to work in segments, rotating the tower and moving the camera down, I also made sure that my images had an overlap of at least 30 %, for the images of the base I worked in square sections like a chest board and worked along it. I used manual focus so that the focal length wouldn't change too much throughout the shooting. 

After this I collected the images and loaded them into agisoft metashape, the first thing to do was create a point cloud which you can see below. 

From this generation of the point cloud, you can see that there are some sections out of place, for the sake of this test I wanted to see what my raw data came out like so I didn't make any adjustments. I decided to create a denser point cloud to pick up more details which you can see below.

I then created a mesh, you can see from the images below of the model that there are areas in the point cloud that need cleaning up, you can also see the sections where there wasn't any image data and that I need to collect images of when I try again. While the base is messed up, the tower itself looks ok, it has some points of information missing but has mostly captured it ok.

While this obviously isn't perfect and needed editing at the point cloud stage, I think that the actual tower has turned out pretty well, it has picked up the details well and from this test, I can see where I need to collect more images.

My next test will be to use the same images but look at how to edit and clean up the point cloud and mesh to get the best result. After this, I will evaluate how the images I collect need to be improved and will retake the photos of the dice tower. I also need to see if there is any way I can collect photos of the water (resin) as this hasn't transferred into the scan at all. I will also look at how I can export the work to then be used in other software, I would eventually like to have a complete scan of the tower, that I can use in after effects of unreal engine and try and put the piece into VR