Examining the First Assignment Brief

Now that I've completed the foundations for my project, it's time to begin working on my first piece of coursework for our first module of the term - Data Capture.

This year our data capture module is focused specifically on photogrammetry. Photogrammetry is the act of taking photographs and turning them into a full 3D model to be used as part of a scene (like how a photograph turns three dimensional reality into 2D, but in reverse).

---

We've been set two items of coursework for the module, the first of which I've included the brief for below:

Data Capture Cswk1 22-23.pdf

Download PDF • 265KB

I feel it's best before jumping into any actual photo taking to breakdown the tasks on the brief and evaluate how I'm going to achieve them. I've therefore copied these into the bullet point list below with accompanying comments:

1. Propose suitable structure(s)/object(s) for scanning (for approval by Tutor). - This actually appears to have changed since the production of this brief, as the final assets we'll be using are captures taken at a trip to Winspit Quarry (to be discussed in a later post).

2. Produce visual concept/plan for the scanning strategy. - Whilst, like I said above, the captures for this particular coursework are already planned out, I can absolutely see the benefit of producing a plan for scanning strategies going forwards using what I've learned as part of this assignment. I intend to summarize my findings and research during this assignment as I experiment with different methods and softwares, but will include a visual aide for the optimum scanning strategy when it becomes apparent.

3. Develop a scanning strategy that most aids an effective image capture. - Hopefully this will also be discovered as part of the above point. Whilst using different methods for captures and conversion, I hope to have a compilation of many different strategies I've applied with which I can draw conclusions from.

4. Ensure that the scans are in the correct file format and that it can be imported into the appropriate conversion software pipeline. - This, theoretically, should happen naturally as the majority of photogrammetry softwares output their meshes into file formats used by industry standard 3D software. However I will be mindful to ensure that all my models are being output as .obj or .fbx files at least, and that they have no issues being used in Unreal Engine, Maya & Zbrush.

5. Prepare all content for Cswk2 that will require you to further develop the assets. - This is something I will obviously go into more detail on when I come to working on the 2nd coursework. But for now, what's important to remember is that these assets will need to stand up to Quixel Megascan assets and be prepared for use in an Unreal Engine 5 environment (including 'cleaning up' the geometry and retopologising them).

As well as the main tasks detailed above, a section of the brief pertains to this blog and the content it needs to show. I've therefore also broken down these points below with observations on how I will achieve what the brief is asking of me:

• Capture Theory - This essentially means that I need to demonstrate an understanding of what's important for captures for the purpose of photogrammetry. Lighting being the big one, but also camera angles, the actual amount of photos necessary etc and how they all impact the resulting mesh.

• Equipment Used - This is quite self explanatory. This simply means what equipment I will be using to obtain my captures. In this case I imagine it will be my phone (a Google Pixel 4a), one of the college's DSLR cameras & the college's iPad with lidar scan facilities.

• Capture Strategy - This is general terminology for showing sufficient understanding of the actual process of capturing assets for photogrammetry. I'll need to ensure my blog displays a clear plan from myself on how I will approach this, with plenty of comparisons and justification for how I came to the conclusion of what process I will follow.

• An Understanding of Subject Limitations - Similar to capture theory, but rather than understanding the process of how to optimize my captures, this pertains to the subject itself. I will need to be aware of what subjects will produce the best results, and what issues I may face depending on factors like; a full 360 degree vision of the object not being available, complex textures, lack of differentiation from the surrounding terrain etc.

• Data Processing - This is a rather ambiguous point, as the processing of the data (captures in this case) will simply be the importing of them from whatever their source may be (lidar scan via iPad, photo on my phone or DSLR SD card) to a PC and add them into my chosen software.

• Software Options, Comparisons and Workflow - I've already selected 3 main softwares I'll be using (Meshroom, Autodesk ReCap and 3DF Zephyr) so this point will be fulfilled when I come to writing comparisons between the 3. Additionally I'll need to justify my reasoning for using each of them, and also write a detailed workflow for each application as well.

• Necessary Asset Clean-Up - Unfortunately, none of the meshes (whether completed via photogrammetry or lidar) will turn out perfectly each time. They will contain far too many polygons with uneven topology and will require 'cleaning up'. I will need to establish what methods I'll be applying to complete this step, again with comparisons and a workflow plan, then optimizing my created assets.

• Evaluation of Results - This is simply drawing conclusions from the experimentation during the assignment. i.e. whatever software is best and why, the same for which is worst, what's the optimal workflow to follow for capturing and asset cleanup and why etc.

That completes my evaluation of the requirements. I now believe I have an understanding of what I need to do to complete this assignment so will get to work experimenting with photogrammetry.

As one final note, the brief does also detail the deliverables we require and mentions a LookDev and breakdown rendered images as one of the criteria. This means I will need to keep a note to produce a short video for my final assets showing the following things:

- Original captures for the model in a brief slideshow (this will demonstrate the origin of the final models).

- Wireframe turnaround render of each final mesh (this will demonstrate the finalized clean topology and polygon count).

- Fully textured turnaround render of each final mesh (this will obviously demonstrate the effectiveness of the final result).

The above is simply a reminder for myself on what I aim to produce for this deliverable, but that brings an end to this examination of the assignment brief!