tl;dr: Gamma Camera is my open source flexible camera array project and you can build your own with https://github.com/robrussell/gamma. I made a thing I call Gamma Camera. What is it? It’s easiest to explain starting from my motivation. For the past few years I’ve been tinkering with a lot of commodity camera parts and single board computers like the Raspberry Pi and Jetson Nano. I’ve also spent a lot of time trying to get a better handle on algorithms like structure from motion (SfM) and, more recently, neural radiance fields (NeRFs).

Here’s another camera mounting block to 3d print. This part is meant to go with my previous design for mounting a Raspberry Pi on a t-slot rig. It’s just the minimum bit of plastic needed to join the camera module v3 to a 1/4-20" nut. Then the nut can be used to attach the camera module to a ball head like this one from Small Rig. Just like the rest of Gamma Camera the idea is to use 3d printed parts to bridge between common COTS parts.

My t-slot obsession has not let up. Last night I made a nice little design to easily secure a Raspberry Pi to a piece of 2020-profile extruded aluminum. This follows on from the 1/4"-20 to t-slot mount that I made recently. Just like that component, I used BOSL2 in OpenSCAD to quickly put it all together. Design Here’s the design for the part if you want to have a look.

Storage prices have dropped drastically since I built my Gamma Camera array. As I mentioned a little while ago, I’m interested in USB storage for the next iteration. In gammacam P6 I have 6 Raspberry Pi 4B nodes. The microSD card holds the OS and also the imagery data. I have heard that microSD card speeds are slower than USB options. And lately I’ve noticed that prices for some pretty fast storage options have been falling.

Three things from this week. This week I’ve got three little progress updates on the next gammacam build. Updating the sensor head I thought I was going to need to change the design of my camera head a bit to fit the Raspberry Pi GS camera module.It turns out that the sensor head will work fine as-is. The plastic cover on the back of the GS camera module is easily removable.

Three things from this week. Raspberry Pis are showing up in stock more often. So I have to get to work and start building the next version of my Gamma Camera rig. I’ve been thinking about it for so long that I’m afraid of falling into second system syndrome. I guess it’d be seventh system syndrome in my case… I’ve got a lot of half-complete thoughts about what I want to put into the next rig so here are three of my best ones.

A Global Shutter Camera for the Raspberry Pi started shipping a few months ago. There have been other global shutter camera modules for the Raspberry Pi but this one comes from Raspberry Pi directly. So it should be well supported by the mainline software including the camera driver and libcamera. Here’s a little summary of some official Raspberry Pi camera modules with the details I care about: Name aka Sensor Pixels Shutter Lens Camera Module 2 v2 IMX219 8M Rolling Built-in Camera Module 3 v3 IMX708 12M Rolling Built-in 75° Camera Module 3 Wide v3 IMX708 12M Rolling Built-in 120° High Quality Camera HQ IMX477 12M Rolling C/CS High Quality Camera HQ IMX477 12M Rolling M12 Global Shutter Camera HQ IMX296 1.

There’s an issue with the way the shutdown command works on gammacam P6. I really like using the little OLED bonnet user interface to control the system and I’d like to fix things so that I can reliably shut it down without needing to check in over ssh. Here’s a quick overview of how the menuing system works. First off, the UI is split into two parts. One file has the hardware-specific part and another that has no hardware dependencies.

Gammacam P6 produces a lot of output files. It takes photos multiple times a second from each camera. Many pipelines don’t need all of the output. In my case I like using Nvidia Instant NGP to make NeRFs from my photos. The first step in processing pictures to use with Instant NGP is to run the SfM (Structure from Motion) pipeline called COLMAP. COLMAP incorporates a number of well-established photogrammetry algorithms to produce useful data from imagery sets.

Overview I wanted to build a self-sufficient portable lab network consisting entirely of Raspberries Pi. The lab network should be able to access the internet if there’s wifi available. And it should be easy to connect an external computer to the network and access any of the Raspberries Pi over the lab network. Basically I’d like to be able to run the whole thing on battery power and use it with or without upstream internet.