top of page
  • Havsfjord

Digital assignment: 3D modeling

On my way back towards Carrboro, we always drive past a little patch of green grass with several statues and installations placed what almost looks like random. Some of them are more complex, moving on their own. When we got the assignment of doing a 3D project in Agisoft Megashape (previously known as Agisoft PhotoScan), I decided instantly I wanted to try to do one of them. JJ advised us when photographing outside to a) not take photos when it was sunny, cloudy weather is preferable and b) do not photograph something too shiny. I had this advice in mind when going out to take my photos, but however in the end I managed to sort of do them both.

All the statues are made by Carrboro metal sculpturist Mike Roig, and the plot of land they’re on is adjacent to his studio and home. I chose the piece Looking up primarily based on its texture being more rugged and not as shiny as the other pieces. On my way to my first attempt the weather shifted pretty fast and I had to do it in sunshine, which meant that the back of the sculpture didn’t come out as well in the photographs. My first attempt I took 42 ( a little overkill) photos, and as you can see from down below, parts of the back didn’t really want to come out.

For my next attempt, the weather was cloudy, but even so the light didn’t really want to come out properly anyway, just as with my first attempt. Even though cloudy I think this one came out even more poorly than my first attempt. Here Metashape told me that out of my 29 photos, four of them could not be properly aligned and that they were cut from the rendering; it didn’t give me more information or clues as to why they couldn’t be aligned (I thought I was going pretty straight and slow but maybe not). My own guess would probably be that it also had to do with the light not coming out properly, making it difficult for Metashape to align them properly. This one however as you can see, did not turn out well as it couldn't properly render the back of it and turned out worse than the first one.

Just for fun, I did a version with all the 71 photos from both sessions that you can find below, hoping it may make it a bit clearer, but I was wrong. As you can see the back has a whole in it that I couldn't properly fill. I was surprised that it turned out well surface wise, as the lightning was very different in the two sets of photos, so the front part of it I would say still looks decent.

Apart from my models not turning out perfect, I thought the software with the workshop and guidelines provided were comprehensible and easy to understand. It was also very fun to see step by step how all these photos slowly turned into an actual model. I could see this being used in numerous ways and would be excited to try it out myself, even though I am not sure how I would use it just yet. I can imagine that trying to troubleshoot or go deeper into the possibilities of the software will require a lot of learning, but by simply following the software step by step (grateful that you can’t accidentally jump over a step in the process for example) and the guidelines is for now enough to continue on doing 3D models.


Recent Posts

See All

Above is timeline about Elsa Schiaparelli; her life, career and legacy, made in TimelineJS. Schiaparelli has always fascinated me (and I’m all but alone in my fascination), as she throughout her caree

As Ted Underwood points out in his text Where to start with text mining.; “Quantitative analysis starts to make things easier only when we start working on a scale where it’s impossible for a human re

bottom of page