Perspective is a bit hard to grasp in this image, but all of this hardware is mounted thirty feet above the ground. This time-lapse photography box makes use of the sun and a Raspberry Pi to document the goings on. The rig is one of three that were built by [Patty Chuck] to record progress on a seventy acre construction site over the course of eighteen months. The gallery linked above shows off the project well, but a much more in-depth text description is found in his Reddit thread.
What’s not shown in the image is a solar array which powers the box. When they were installed there were no utilities on site. To guard against power-loss there’s a hardware RTC that keeps ticking. The Raspberry Pi uses GPIO pins to switch the Nikon D7100 camera on once every five minutes during the work day. It snaps a photo before powering it down again. It also monitors a temperature sensor and actuates circulation fans if necessary.
He’s planning to post the videos once the project’s done in 18 months. If you see them and remember this post, send us the link and we’ll post the update.
Very nice machine, but please adjust the horizon.
Luckily, a postscript can fix that for a tiny drop in field of view. With such high resolution stills, he’d still be downsizing even for 1080p, so quality shouldn’t be an issue. Yay image processing!
-Dude, I became a dentist!
-You got a degree?
-No, I bought a dental drill.
Hmmm, that would be nearly 35 000 power cycles, can a camera like that really handle that many on/off cycles?
why wouldn’t it be capable of doing that 35 thousand isn’t a lot of clicks
Starting up DC/DC converters wears them out faster than running them all the time, the start current is higher than steady current since you are charging capacitors in the whole circuit.
Also the camera will probably run the anti-dust system and other things every startup and those systems may not be rated to be used that often.
The shutter is only rated for 100-200k operations so i wouldn’t think they would design the rest of the camera to last as many as 35k power cycles.
Startup antidust (or in Olympus’s case at shutdown) can usually be disabled in menus. That said there’s many other components that can stress when you actuate the shutter. I don’t buy that the load on a capacitor or inductor will wear them out after 35k cycles. In the case of many switchmode supplies these things are designed to handle a continuous form of ripple. They aren’t indestructible but they should last a hell of a lot longer than 35k cycles.
why not use the rpi to transfer the picture to some storage server that way you have a use for the rpi and you dont have to replace the SD card
never mind i read the indept article, quite nicely done
To be fair he is using a wifi dongle to run a ad-hoc wifi network so he can pull photos down via SSH/SFTP. Perhaps some reading is in order?
Details of Raspberry Pi: The Pi is running your basic install of wheezy, but I put it on a 64 GB SD card. The Pi also has, hooked up to the GPIO, a real-time hardware clock (in case of temporary loss of power), and a circuit board I designed that controls both turning on and off the camera (the camera is not powered on between shots to conserve precious electrons) and to control two circulating fans inside the box. There is a temp probe hooked up to the GPIO, and if the temp gets above 90° F, the fans kick on. The Pi also has a wireless USB fob and is set up as a wireless access point. That way, I can sit on the ground at the base of the pole and SSH to the Pi to either change camera settings or download (via SFTP) images saved to the SD card. (We had debated tossing a MiFi in the box, but figured over the course of 18 months with three system, that’d be about $11,000 in lining Verizon’s pockets.) For camera control, I’m using gphoto2, which is a very powerful command-line based camera controller. Everything is run using crontab (hence the importance of a real-time hardware clock). The cron job runs a little python script every three minutes that checks the temp/controls the fans, and the cron job also runs a simple shell script every five minutes between 7am and 4:30pm Monday-Friday that turns on the camera, takes a photo, saves it to the Pi SD card, and then shuts off the camera. There’s also a cron job that runs every minute that logs the temp and whether or not the fans are on (only so I can get an idea of how hot it’s getting in the box and whether or not the fans make a difference. So far, looks like internal temp is running about 25% hotter than outside ambient temp, but by simply circulating the air in the box, I can lower the internal temp about 10%!)
Please tell me that the opening for the camera in the front of the box isn’t just a hole? It sure looks like it but I have to be missing something…
“The whole front of the box is a plexi window. I covered the majority of it with aluminum foil except where the lens is located. My hope would be that a majority of the solar heat is reflected instead of absorbed in the box.” Yaay for reading. :P
Might help to put the foil on the OUTSIDE of the case. As this is, AIUI, light’s going to come thru the foil, be turned into heat, which then won’t reflect off the foil. Meanwhile all the heat radiated from inside the box is gonna be reflected back.
I could be wrong, I’m going on my memory of how greenhouses work. But I think this is slightly worse than not bothering. I think it’d be better to just paint the whole outside of the box white. And are there vents in there? Or just fans to blow air round in circles? A simple “chimney” with a conical “hat” on to keep rain out would seem to do. Indeed a chimney can actually suck air through as wind blows past it.
Cool project. Looking forward to seeing the end result.
Most SLR cameras like that consume next to zero power when in standby… Considering you have a 2+ watt Raspberry Pi sitting there on all the time, I wonder if its worth turning the camera off.
Also, do test this system over an extended period before you use it for real – raspberry Pi’s, in my experience, freeze after a couple of days or weeks for no reason I can identify – my theory is either a kernel memory leak or something up with the GPU firmware that halts the CPU.
the raspi i am using has a nas + torrent box has an uptime of 55 days, never seen it fail so it was probably just some mantainance i did that required a restart, these things are rock solid if you get a good psu
Strange. I have had weeks and months of uptime on my RPi xbmc. Which reminds me I should reboot it so it downloads some updates.
I run a Raspi in a remote weather station which is networked via GSM and VPN so there is no need for the local Ethernet adapter to be enabled. I used to leave it enabled and disconnected because I didn’t think it was a big deal, but the thing would crash every 1 or 2 days. After turning the Ethernet port off (if-down eth0) the thing runs days on end without any problems, I think the Ethernet drivers or chipset are buggy somehow.
I wondered where I could pick up a D7100 for cheap….
Where was this building site again?
That’s the advantage of the 30 foot pole.
And a bunch of burly builders. From a distance you don’t know there’s a nice Nikon in there. And it’s only a particularly stupid thief who tries to steal a security camera. Tho I’ve known it happen, once.
I’d imagine the market for stolen security cameras is only a bit more lively than for stolen burglar alarms.
Boring, We must always use a RasPi or Ard…when a small Pic8 or Avr 8 pin part will do. How about learning to program an wmbedded processor(or is that too much work).
have you noticed the price for a wifi controler vs a rapi + wifi dongle?
counting the development time you could be talking hundreds of dolars in firmware and testing vs cots hw and sw and someone’s python scripts
An interesting project shouldn’t be passed over just because it uses something that’s popular. If you want to see an ATtiny, a PIC, or whatever else, then go do a project that uses your favorite microcontroller, and submit it. If it has merit, it will likely be posted.
Otherwise, quit your bitching.
I like how hackaday seems to be posting enough about a project to give people like you something to bitch about, using buzzwords like Arduino or Raspberry pi to give you autists a heart attack. I admit there is a lot of things that don’t warrant a pi or arduino but if you read the source you would realize that what he is doing is all but beyond the scope of what most microcontrollers can do.
Onoes! Some guy used an RPi to develop a nifty timelapse camera setup instead of going to engineering school for four years so he could spend even more time developing his own board!
Wastefulness is not an issue when it comes to one-off projects, be they prototypes or hobby projects. Ease of development, total price (including time spent), and convenience are the major factors here. Is that really so goddamn hard to grasp?
And even with large-scale production, sometimes ease of development and convenience edges out leaner solutions in terms of cost (or even development time — sometimes it’s better to just get a product out the door quicker). That goes doubly so if a product is meant to appeal to DIYers, modders, and general hobbyists.
Hi, my setup looks similar, except I had a challenge to adapt point & shoot camera. Take a look here http://lukse.lt/uzrasai/2013-07-ilgo-laikotarpio-time-lapse-fotografavimas/
In the event someone criminally bent who owns lineman’s equipment who is nearby, and recognizes the location from the photos, could have a nice camera before construction starts. Would anyone(other than Patty Chuck) suspect that a person capable of climbing the pole isn’t authorized to “service” that equipment? Any hacker who is in decent physical shape, and is any sort of hackgineer at all can should be able to build something to get to that equipment