Like Photography years ago, 3D Scanning — or, more broadly known as Reality Capture — has become affordable for consumers and small businesses. While professional grade 3D scanners are still expensive, two technologies that are commonly used for 3D scanning on a budget are Depth Sensor based 3D Scanning and Photogrammetry, which uses a regular camera and special software.
But like Photography, buying just a scanning device and software might be enough for hobbyists but won’t give the best results. If you want to make 3D scans for professional purposes, you need to invest in some extra equipment to populate your (mobile) 3D scanning studio. Luckily, these can be found for prices that match the affordable scanning hardware and software.
My definition of affordable for this post is using hardware and software that won’t exceed an investment of $1000 (not counting a computer of tablet you probably already have).
Further down this post you’ll find a selection of essential products that I actually use myself to make many of the 3D captures you see on this website. And I’ve taken the time to search for the best deals, so you don’t have to.
But if you haven’t done so already, you should first choose which 3D capture technology you’re going to use.
Depth Sensor Scanning
Using a depth sensor or going for photogrammetry is not an easy choice to make. In fact, I advice to get experience with both approaches so you’ll get a good feel of which technique is best suitable for each project you encounter. To help you, I’ve summarized the differences between the two technologies:
Using a Depth Sensor for 3D Scanning
First of all, let’s take a look at what it looks like in reality and what you’ll see in a typical depth sensor scanning program:
In my opinion, sensor scanning is what you should start with if you’re totally new to 3D scanning and haven’t got a lot of experience with technology in general or photography. You won’t get the very best scan quality but the results are light-weight (small file size and low memory usage). This makes them suitable for many (realtime) 3D purposes. The hardware is fool proof and the software is very user friendly. You don’t have to worry about the background behind your object and it will give an interactive preview of the scanning progress. Scanning is fast (under two minutes for a typical session) which makes it the best way to capture people on a budget. Processing is also fast so don’t need a very fast computer or tablet.
Typical uses of depth sensor scanning are capturing people and body parts (for both aesthetic and measurement purposes), objects to use as low-poly assets in games and VR/AR experiences (for reference or actual use), rooms (measurements only).
Below is the result from the Sense 2 from the video above. Click on the gear icon and select Rendering > MatCap to view the geometry without textures.
Which Depth Sensor should I buy?
You can start experimenting with Photogrammetry if you have some experience with photography and manual camera settings (even if you plan to use a smartphone) — and don’t mind that your first attempts will fail. There’s a learning curve but once you gain some experience, photogrammetry can generate very impressively detailed 3D models — both for textures and geometry. The object size and capture range is merely limited by what’s physically possible to capture. With the right camera and lens you can capture anything from a toy to a large statue. By using a camera drone you can even capture entire buildings.
Photogrammetry is not the holy grail of 3D capture and is limited to objects (or environments) that have a lot of visual detail and it doesn’t like movement — either of the subject or the background. This means that single-camera Photogrammetry is not a great idea for capturing people since they won’t be able to stand totally still long enough. It’s also not suitable for objects with a lot of large, flat surfaces, like furniture.
Typical uses of single-camera Photogrammetry are capturing statues of all sizes (cultural heritage, toys), organic objects (archaeology, nature documentation), high-quality 3D asset creation for games and VFX, areal surveying.
Below is the result from the photo shoot above, processed with the Free version of ReMake (standard Quality). Again, click on the gear icon and select Rendering > MatCap to view the geometry without textures.
Which Photogrammetry Software should I use?
Since this post is about affordability, let’s with three FREE solutions, all capped at a maximum of 50 photos which is just enough for most small to medium objects:
- Autodesk ReCap 360 (Review) offers a free version. It’s cloud based and works through a webbrowser interface. This means you can use it on any computer (Windows, Mac, Linux or even an Android smartphone) and it doesn’t even have to be a fast one since the processing is done on on an external server. Unfortunately, it offers no post-processing and clean-up tools.
- Autodesk ReMake (Review) is available as a free version that uses the same cloud processing engine as ReCap 360 but in the form of a Windows application that does offer a wide array of post-processing tools to retouch and optimize the 3D model. If you want to process more photos, locally a computer with a fast Nvidia graphics card, and with better (Ultra) quality, you’ll need to update to the Pro version for $30 monthly or $300 annually in the US and €35 monthly or €350 annually in Europe.
- 3DF Zephyr (Review Soon) recently released a free version. It’s a Windows application that offers only local processing. It can run on a mid-range PC but a fast computer with modern Nvdia graphics card will speed up processing significantly. If you want to process up to 500 photos you can get an the Lite version for a one-time price of $149. Unlike ReMake, Zephyr’s free version has no quality limitations.
If you don’t mind paying for software, you can also get:
- Agisoft PhotoScan (Review) is available for Windows, Mac and Linux and can benefit from all kinds of graphics cards that support OpenCL. The $179 Standard version can process an unlimited amount of photos but the actual amount greatly depends on the amount of RAM in the computer.
- Capturing Reality RealityCapture (Review) is available for Windows and surprises with its incredible speed. Processing that takes hours in other programs can be done in minutes with RealityCapture. It’s not as RAM-hungry as PhotoScan but strictly requires an Nvidia GPU. It has a subscription-based business model for $40 a month or $99 for a 3-month license (capped at 2500 photos).
More Detailed Pros and Cons for if you’re not in a hurry
In my opinion, using a depth sensor is the easiest way to start 3D scanning. It requires very little technical knowledge because the process is very interactive. Below you can see an example of what sensor scanning looks like in reality and what you’ll see on the screen of your laptop or tablet.
+ No need to worry about the background
What’s great about a depth sensor is that it actually senses depth. That means that the software can tell it to only capture the object or person — and ignore the background. Because of this, the background doesn’t matter at all. There can even be people moving in the background.
+ Interactive Progress Preview
All depth sensor software overlays a preview of what is being captured on top of the camera feed. This way you know exactly which parts have been scanned to ensure you cover everything to generate a model without holes (extra important for 3D printing).
+ Fast results
Scanning an object or person can usually be done within two minutes. Processing time depends on the software and computer that’s being used but is usually done within minutes.
+ To Scale by default
Depth Sensors are factory-calibrated and know exactly how far objects are. Sensor scans are usually great for doing interactive measurements that don’t require industrial (sub-millimeter) accuracy.
– Don’t work well outdoors
Most depth sensors use a small infrared laser projector. This light is invisible to the human eye, which is convenient and safe. But the small projector isn’t as powerful as other infrared light sources like, let’s say… the sun. You might be able to capture on a very overcast day but any direct sunlight messes up the depth sensing capabilities.
– Limited Scan Range & Object Size
There are different kinds of depth sensors but the ones I’m covering here are so called short-range sensors. This means they can sense objects from a distance of 20-40 cm up to 100-300 meters depending on the exact device. This limits them to medium to large objects and people. I’d say anything from 40 – 250 cm tall is the ideal scan size for depth sensor subjects.
– Limited Scan Quality
I determine scan quality by both looking the geometric details and texture quality. These depend largely on the resolution of the cameras that are used. Without becoming too technical: these sensors use an infrared black-and-white camera to capture the patterns projected by the infrared projector and and a RGB color camera to capture texture information. These are getting better with each new generation of sensors but will never be as good as a dedicated digital camera that’s used for photogrammetry.
The quality of a depth sensor scan can be considered a low-poly 3D model. It might not be the best scan result but the fact that it’s “light” means it’s perfect for realtime 3D purposes such as online sharing on Sketchfab, VR/AR applications, body measurements and small-scale decorative 3D printing (e.g. miniatures in full color sandstone).
+ Size doesn’t matter
Depending on the camera you use, you can capture objects as small as a toy and as big as a statue. You can also take photos with a drone to capture really big things. The only thing you should be careful with is the depth of field when capturing really small things. You basically want everything in focus, which is impossible with macro photography. DOF bracketing might be an option here.
+ The geometry quality can be incredible
When using a camera with a sharp lens and many megapixels, Photogrammetry can achieve incredibly detailed results that can match professional 3D scanners that cost upwards of $15,000. But this really depends on the visual details of the object you’re scanning.
+ The texture quality is always incredible
What always sets photogrammetry apart from even the most expensive industrial 3D scanners is texture quality. This can make up for less detailed geometry and can also be used to generate special texture maps like normal maps and displacement maps that can enhance the realism of 3D models without adding extra geometry.
– It requires experience
I get a lot of emails from people that loaded a few (I’ve read stories of just 4) photos into Photogrammetry software and complained that the results where lackluster. To make photogrammetry work, you’ll first have to understand how it works and how you can take the photos in a way that’s optimal for a computer to understand. That often has little to do with human logic. Even after hundreds of photo shoots and testing almost every software suite available, I’m still improving the quality of my photo sets with each new project. The problem is you have no zero feedback while taking pictures and not all problems that arise in processing can easily be fixed. I’m working on a complete list of shooting tips.
But the experience is not just in shooting the photos. To get the best quality out of Photogrammetry software, you’ll also have to learn what impact certain settings have on the quality/processing time balance. That’s significant if you want to offer it as a professional service or have limited computer resources.
– It’s slow
While using a multi-camera Photogrammetry rig is the fastest way to 3D capture objects and people, doing it with a single camera requires time and patience. Shooting from a tripod is a must and to get the best shots you’ll sometimes use longer exposure times. And even for small to medium objects, you will have to shoot a minimum of 40-50 photos from various angles to get good results.
Processing the photos can also take a long time. Greatly depending on the amount and size of the photos and the specs of your computer, it’s not unusual to have to wait hours before getting results. Some programs require so many resources that you cannot do anything else on the computer in the meantime. It’s good to know that many modern Photogrammetry programs are optimized for GPU-acceleration. In most cases, one or multiple graphics cards can be used to speed up the algorithms significantly. And in most cases, those from Nvidia are the best (or only) option.
The problem of speeds wil become less of a problem in the future because algorithms are getting more efficient and computing hardware is becoming faster and more affordable. Both as it stands now: patience is a virtue.
– Motion is Killing
While most photogrammetry algorithms can handle moving objects in the far background, they can really go haywire when larger objects or people move even a little bit. But keep in mind that in difficult cases, you can always mask out distracting elements before processing. But again, this takes time.
– It’s not to scale by default
Most photogrammetry programs try to scale the model based on camera and lens information but this is always an estimate. So it’s always a good idea to take some reference measurements of the objects your’e capturing or put something in the scene of known size to scale the model later.
Essential Stuff for your
3D Scanning Studio
Now that you’ve made up your mind about using a sensor or photogrammetry, let me guide you through all the other hardware that will make the difference between a hobbyist’s scan and one that looks great!
1. Light, Light & Light
While industrial-grade 3D Scanners usually have a build in LED light source, affordable infrared depth sensors haven’t. This means the scan quality will greatly depend on the lighting conditions. For photogrammetry the solution is simple: capture your subject outdoors, preferably on an overcast day for nice, diffuse lighting. If you have a 3D scanner that projects a visible line of laser light, scanning outside might also work. But if your scanner uses invisible, infrared light, you’re bound to scan indoors.
But in many cases you’ll want to scan indoors anyway to keep control of the lighting. In this case it’s wise to invest in a studio light setup. Not flash lights for photography, but permanent lights used for video production. The kits usually come with 3 lights with softboxes on stands. Depending on your budget, you can buy a kit with LED lights or a more affordable version with fluorescent bulbs. The amount of light you need depends on what you want to capture. Just be sure the light temperature is 5500 Kelvin (daylight).
I have three ways to light my 3D scan subjects:
Small Light Kit
This is a very portable (they usually come with a carrying bag) kit with two or three small, single-bulb, umbrella-style softboxes. It’s good for scanning small objects like the Teddy Bear you see a lot on my blog.
They’re usually just standing in a corner of my office and I use them for most of the photography on this site as well.
Large Light Kit
I also have a larger that can hold 5 separately switchable bulbs, totaling 2500 Watts of light. It comes with a bag, but I call it the Body Bag for a reason. That’s why I only use this set in my studio only and usually for Full Body Scanning.
I might swap that kit for an LED version one day, because that would be a lot lighter and would save a lot of electricity as well. You can see my kit in action in the animated GIF in the Turntable section of this post.
Mobile LED Light
The LED mounted to the headphone jack of an iPad for 3D scanning with Structure Sensor (Review).
The LED mounted to the Tripod Mount of the Sense 2 3D Scanner (Review) with the included adapter.
2. A Turntable
Using a turntable has the big advantage of not having to walk around the subject. This means you’ll need less space and less lights and in some cases your 3D scanner can even be mounted in a fixed position.
Small Manual Turntable (for Depth Sensor)
You can use a turntable with a scanner on a tripod or handheld. Just make sure your hands and arms are out of frame as much as possible so they won’t mess up the tracking.
A Smart, Automatic Turntable (for Photogrammetry)
A Large, Automatic Turntable (for Depth Sensor)
If you want to scan bigger objects or humans, you’ll have to buy something more serious. I own an electric turntable manufactured by Dutch company Pre-Motion, that can hold up to 100 kilograms. They have two versions with different fixed speeds and I tested both to discover that the slowest 0.8 RPM version works great for 3D Scanning. I modded mine by adding a simple foot-operated light switch in the power cord, so I can take my time to scan the front and head of people first and then turn on the turntable with my foot to capture the rest at a comfortable pace (the GIF below is played at 4x the original speed). A turntable like this isn’t very cheap and also isn’t very light (15.5 kilograms), but if you’re going to put people on something that rotates it better be good, right?
To stitch geometry from different angles together into one seamless model, all 3D capturing software use feature detection and tracking to understand which parts need to go where. Some 3D scan solutions use just the geometric features (from the black-and-white Depth Map) for this, while others also use the color information for this (referred to as RGB+Depth tracking). This means that the former method will not be able to capture perfectly smooth globe because of the uniform geometry of a sphere, but the latter can, because it detects unique features of the print.
4. Matting Spray
Both 3D scanning hardware and Photogrammetry-based software solutions rely on computer vision of some kind. They either project visible or invisible light onto an object and capture the result with a camera, or they use software algorithms to estimate depth from images. In all cases, the computer tries to understand what it’s seeing. And computers aren’t smart enough yet to understand a few surface types that humans take for granted: transparency and specularity.
If your want to scan people, you will run into problems with people wearing shiny things like glasses. If you need a 3D model of a person that always wears glasses, you can scan them without their spectacles on and add these in post production. For this you will naturally need a 3D model of the glasses, and 3D scanning them separately is very hard. Luckily, you can find many brands and models on 3D Stock websites like Turbosquid for a reasonable price. Just be sure the file format is compatible with the 3D modeling software you intend to use.
5. A Sketchfab Account
If you want to share your 3D Scans with the world (and don’t have the cash to 3D print all your scans and mail them to your friends and family), there’s only one way to do it in style: Sketchfab. The website has quickly become the YouTube of 3D Models and offers an easy-to-use way to publish and share your 3D scans through the web, social media and even in Virtual Reality (I wrote a tutorial about this). Manually uploading scans isn’t necessary in most cases, because almost every 3D scanning and photogrammetry software can directly publish the result to your Sketchfab account.
It’s free, but I can totally recommend the Pro account if you’re taking yourself and your work seriously (and want to support the development of this great service). The Pro features I use the most (Yes! You can also follow me on Sketchfab) are the ability to publish models privately and change the background to something more branded. Also, some of my scans go beyond the 50MB file size limit of the Free version.
Thanks for reading. I hope this post is useful for you if you’re getting into 3D Scanning (which you should). If you think it would be helpful for your friends and followers, please use one of the buttons to share it on your favorite social network. For this you can use the buttons on the left (desktop) or under (mobile) this post.