How to design an enclosure using photogrammetry? – The Complete Guide

In this article, I’m gonna show you step by step how to design an enclosure just like this with the help of modeling clay and Photogrammetry.

Lately, I needed an NRF24l chip-based RC Transmitter for my project. I could go into a classic Playstation, or X-Box pad look. But I decided to design something ergonomic and perfectly suited to my hand. So I did it. And now, I’m gonna use it as an example to tell you how you can use photogrammetry to design these kinds of organic-looking enclosures for any device.

All the software we’re gonna use today is free of charge. So all you need is a camera (the one on your phone will be sufficient), a PC, and access to a 3D printer.

However, if you don’t want to mess with photogrammetry, you can download the .step file from HERE, and go straight to editing in e.g. Fusion 360. Note that I was making this model for myself. So if you have a very different hand than mine, it may not suit you so perfectly 🙂 It’s best if you print it out and check it yourself.

Video version

What’s Photogrammetry?

Photogrammetry is a method of creating 3D models of physical objects from photographic images. This isn’t a new technology. In fact, it’s pretty old. The first documented use dates back to 1858 in Paris. Currently, it’s used, in geology for the search for minerals, in architecture for the inventory of monuments, as well as in archeology, meteorology, and cartography.

Backstory

If you search the internet right now, you’re gonna find plenty of articles and videos about photogrammetry. Today’s technology has advanced so well that it can be successfully done at home. You don’t need to have a computer from NASA. 

However, most people use it to model statues, rocks, trees, or them selfs. 

And there’s nothing wrong with that. I started by creating a model of an old tombstone that I found not far from where I live. It turned out really nice, but otherwise, it’s actually.. quite useless. What can I do about it other than show my friends what a superb model I made? Not much.

Then I came up with the really neat idea that I could use this technique to design an enclosure for my RC transmitter. Instead of modeling a solid body in the classic way, I can use modeling clay and then transfer it to a computer thanks to photogrammetry. It’s much easier and faster to get a shape that perfectly fits my hand this way. 

Plus, it’s gonna looks awesome and unique as hell.

Okay, there’s a lot to cover, so let’s get started!

Step 1. Sketch a preliminary model on a piece of paper.

Before you even start, you should have the overall concept in mind. Just put it on a piece of paper. Seriously, don’t skip this step. This is a good moment to consider what you really want from your device. What functions should it have and what components it must contain.

In my case, it’s gonna be a joystick (no. 1), two potentiometers (no. 2 and 3) for acceleration and calibration, and a battery box (no. 4).

The big things are the most important because you have to estimate how much space you need. These little ones should fit in between. 

Step 2. Create all the necessary components in Fusion 360.

Create a new project in Fusion 360, and add to it all the components you defined in the previous step. Now, it’s good if you know exactly what you want to use. Check your favorite electronic store (local or online), what is available and what best suits your project. If you’re going to use common elements, most likely, you’ll find ready-made models on grabcad.com. If not, you have to draw it yourself.

Step 3. Arrange these components in space.

Following your paper sketch, place them where you want them to be on your device.

I’m gonna place the joystick somewhere in the front – for easy access with my thumb. The first (trigger) potentiometer under the joystick. The second one on the back of the device, because I’m not going to use it very often. And the battery box somewhere in the middle.

Do it roughly. You’ll determine the exact position later.

Step 4. Draw a simple box around all of the components.

Enclose all components in a super simple box. It’ll ensure enough space for everything you want to fit inside after creating the enclosure.

Step 5. Print that box.

Print that box out, or have someone else print it for you. This will be the core of your clay model.

Step 6. Design your enclosure from modeling clay.

This is definitely the weirdest step, especially for tech guys like you and me. Take the modeling clay (it doesn’t matter what kind, generally, the cheaper, the better) and stick it around the box that you’ve just printed. 

Take your time. Make sure everything fits as it should. And that it looks the way you want it.

I needed an hour for this. But if you are more artistically gifted than I’m, it’ll probably take you less.

When you’re satisfied, let it cool down for a while.

Step 7. Add some texture.

Photogrammetry doesn’t like smooth and uniform objects. So you need to add some texture to your model. I’ve tried different approaches, from mixing modeling clay with sawdust to spraying it with colored paint. And by far, the easiest way is to use a marker pen. Of course, the color must stand out. Make as many dashes and dots as you like. Don’t hold back.

Step 8. (optional) Prepare a white sheet of paper on which to put your model.

On this white sheet (just like on the model), add some kind of texture. This piece of paper has three functions:

  1. It has a pattern that will make it easier to calculate the position of the camera.
  2. It reflects some light, illuminating the model from below.
  3. It has a known dimension, so it will make it easier to scale the model afterward.

Step 9. Put the model in an easily accessible place.

You’ve to take a lot of photos on each side of the model. Therefore, the model should be placed somewhere there’s enough space around.

You can put it a little higher, so you have better access from below. A stool in the center of the room would be a good choice. But use whatever you’ve at hand.

I’ll use an old lampstand with a piece of the board instead of a lamp.

Step 10. Provide good lighting.

This is the most crucial element. Without proper light, you won’t get good (if any) results.

It should be soft, uniform, and unchanged when taking photos.

The best light for photogrammetry is outdoors on a cloudy day. But if that’s not an option for you, don’t worry, you can do it at home too.

In that case: 

Use more than one light source (preferably at least three). Additionally, put a piece of white material between the model and the lamp. It’ll diffuse the light nicely. You can also reflect light off a wall or ceiling, which will increase the surface area and thus also soften the light.

Make sure that you don’t cast a shadow on the model yourself as you walk around and take pictures.

Step 11. Provide a good environment.

This step is straightforward, and it comes down to: don’t clean the room 🙂 The more things are in the background, the better. Single-color walls with no reference points will make the photogrammetry process very difficult. 

The second thing is those background items shouldn’t move. Pay attention to this, especially if you are taking photos outside on a windy day.

Step 12. Set the camera to manual mode.

A mid-range smartphone is enough to follow these steps. But if you have a DSLR or a mirrorless camera, that’s even better. 

Since I don’t have such fancy gear, I’ll use my smartphone.

Continuing the topic from the previous steps, you need to keep conditions as stable as possible. So what if you provided the perfect lighting and environment when your camera changes exposure between shots. It can also ruin your results. 

This is why you should be able to access the manual modes of your camera. If the app on your phone doesn’t allow you to do this, you can always install another one.

The most important is to block exposure and white balance. Set some values before the first photo and lock them.

Step 13. Take about 50 photos of your model.

Theoretically, the more photos you take, the better. However… it’ll extend the already long process of creating the mesh model. 

I tried different options, from 20 photos to over 300. And about 50 – 60 shots, I think, it is the most reasonable. Increasing this number won’t significantly improve the quality.

Additionally, if you want to speed things up, you can reduce the resolution to FullHD.

Ok, but how should you take these photos?

Make about three laps around your model, taking a shot every bit. With each subsequent lap, increase the height of the camera.

The general rule is that each photo should overlap around 70% of the previous one. But don’t take it too literally. If you have a model of a similar size as mine, take 15 – 25 photos per lap, and it’ll be ok.

Step 14. Make sure all photos are ok.

Before proceeding, make sure that the photos you took are ok. All images should be a sharp focus and similar to others. If any photo stands out, delete it. Better fewer photos of good quality than more bad ones.

Step 15. Upload all photos to Meshroom.

To add photos to your project, just drag and drop them. Or go to FILE -> IMPORT IMAGES.

Step 16. Add information about your camera’s sensor size (if needed).

Meshroom needs hardware information about the camera to do the calculations correctly. 

If you see a green circle next to photos, you can skip this step. This means that all the necessary information is contained in the photos file.

If it is yellow, it means that some information is missing. Most often, it’ll be the size of the camera sensor. Meshroom will try to estimate this value but most likely will do it wrong. Therefore, I recommend you add it manually.

To do this, you need to find out what sensor size your camera has. The easiest way is to search the Internet or install an application like “detect my hardware”.

Then go to “cameraSensors.db” file (in Meshroom catalog) and add your device. I’d to add my Samsung SM-G973F phone. And then reset Meshroom.

If everything went well, you should see green circles 🙂

Step 17. Start calculating.

While it may seem tempting to click that big green button at the top, don’t do it! Trust me, especially at the beginning of your adventure with photogrammetry.

Instead, go to the nodes below. If they seem complicated to you, then you are right. There are a lot (a looot) of options out there. Fortunately, the default settings are ok in most cases, and you don’t need to change anything.

Go to the “StructureFormMotion” node, right-click, and then “compute“.

Step 18. Wait

I don’t know what to write here. You’ve a short break.

Step 19. Evaluate the initial result.

Up to this point, the process goes relatively fast. And now, you can determine if something is going terribly wrong.

In the 3D viewer, you’re gonna see the initial effects of your scan and all camera positions.

If something looks wrong at this stage, e.g., cameras are at some strange angles, stop the process. It would be a waste of time. Better make sure everything is ok with the photos you uploaded.

On the other hand, if everything seems okay, you can continue. Right-click on the “Texturing” node and “compute”.

Step 20. Wait some more

This time the break will be longer. You can go out for lunch 🙂

Step 21. (optional) Clean and repair the mesh file in Mesxmixer.

This is an optional but highly recommended step. The generated Mesh will have several hundred thousand faces. Fusion 360 really doesn’t like such objects.

That’s why I recommend you to use Meshmixer first. It’s also Autodesk software. Works much faster, and it’s completely free.

In Mesmixer, you can clean, repair, and smooth your Mesh. You can do a lot more, but the rest will be more convenient to do in Fusion.

I know the textures look weird (I don’t know what it depends on), but it doesn’t matter anyway. Start by turning it off completely.

Cut out all the rubbish around the model. Leave a sheet of paper. You’re gonna need it to scale the model.

Now you can perfectly see all the imperfections of the mesh. But don’t worry, it’s fixable, and I’m gonna show you how simple it is. 

  • Start by turning on the Wireframe. Now you can see how many faces your model has. There’s a lot of it, isn’t it?
  • From the menu on the left, find the Select button.
  • Adjust the brush size to your needs.
  • Highlight the defect and a little around it, and DELETE it.
  • Now, from the menu select: Edit -> Erase & Fill.
  • Tweak the Scale and Bulge settings to match the way the hole is filled.
  • When you are satisfied, click Accept.

Repeat the steps for all defects in your model. It can be a bit of work, but it’s worth your effort. The final model will look awesome!

As the final touch, you can globally smooth the entire model.

  • Go to the Select button again.
  • Double-click on a model to select all.
  • From the menu, select: Deform-> Smooth.
  • Play around with the settings and click Accept when you are satisfied.

Depending on how your scan came out, this process may take some time. But take the time and do it thoroughly. This pays off later when printing.

Step 22. Import the Mesh file into Fusion 360

To import a mesh into Fusion 360, create a new document. You could use the file you’ve created instead. However, to modify Mesh, you need to disable design capture history. Which is super helpful in the classic design method. Therefore, it is better to create a separate file and then import the finished effect to the main document.

So, after creating a new document, right-click on the project name and find: “do not capture Design History

Then go to preferences -> Preview Features, and make sure that the Mesh Workspace checkbox is selected.

Now, you should see that the Mesh tab has appeared. But before you go there, import your fixed and smoothed model. Insert -> Insert Mesh

Step 23. Scale, Reduce and transform it to Solid.

Now, go to the Mesh workspace. Before you convert this Mesh to BRap (a solid body), you should do three things.

  • Orient the model well in space. 
  • Scale it to the actual dimensions. This is why you used a sheet of paper under the clay model. If you ensure that the sheet will have the right size, the rest of the model will also be ok.

Cut off the part you no longer need using the plane cut tool. As a Fill Type, select minimal. Thanks to this, you will get an (almost) flat surface on which you will print later.

  • Reduce the number of faces. The more you reduce, the less detail the model will have. The upper limit allowed by Fusion is around 50k. But I made it a lot smaller anyway because my computer isn’t the newest anymore 🙂

Now, you have a fully editable solid body. Think how long it would take you if you wanted to design it with the classical method. Of course, assuming that it is even possible.

Summary

And that would be it when it comes to using photogrammetry to create the device housing. Now it’s time for classic modeling in Fusion 360. With a solid body like this, you can do literally anything. Slice, cut, add, extrude, subtract, and whatever you want.

The next step will be to insert the electronics and all the mechanical components from the second step. But this is a topic for a separate article. It’ll appear in the “projects” section of my website, so stay tuned!

Related Articles

Scroll to Top

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close