| Date | Tasks |
|---|---|
| 7/7 | Wrote my report for last week. I spent the rest of the day reading and watching videos on how to improve my photogrammetry models. |
| 7/8 | I spent all day looking into how to create better photogrammetry models. |
| 7/11 | Created a model with pictures from last week. I tried to play around with control points, and got a good combined photogrammetry model. |
| 7/12 | I started playing with RC on craeting high detail meshes. I got a mesh that was not much different than before; however, it seems that it might yield better results. Went to Oldenborg to take 175 images. |
| 7/13 | Worked on craeting the high detail mesh and model. I had some trouble with space, so I am pausing that to clean up the disk a bit and get the computer ready to handle high detail. I need to detail several 100s of GB of data from the past couple of weeks. |
- Photogrammetry:
- Created the normal detail photogrammetry model with 650 images. 570/650 were recognized, better from last week. Got this by playing with control point.
- However, it seems, like the union of these models is a bit choppy, the floor seemed a bit broken in the union points.
- Working with high detail meshes, hoping for better results.
- I have been reading and watching videos on creating better results, better images seem to be the best bet, but also the sharpening on the model.
- The models from RC, although better from before, are still noisy and not as sharp as we wish them to be. I plan to read more on why there is noise and how to sharpen these models so that we start getting more realistic models.
- I plan to keep working on my skills to create better photogrammetry models.
- The difference from last week is that I have an idea of where to go from now on.
- Work with the tools that RC has to create better models, that includes control points, merge components, unwrapping, and also cleaning up the models.
- Work with ZBrush to clean up the models and get sharper results. I need to get a trial license for this one.
- Keep on perfecting the models.
- Try creating another scan with LiDAR.