Automated Timelapse: 2019 Update

I like this time of year, a chance to reflect on the last 12 months and take stock of accomplishments and realise the achievements. And something I like to gauge a success on is the longevity of a solution, and a time-lapse comparison 6 months apart is seemingly my go to example.

To elaborate on this achievement, earlier this year was the setup of a homebrew CCTV solution using an array of Raspberry Pi’s with cameras, and a VM Cent OS server acting as a PVR host. A surplus Pi W Zero was pointed at the hills and used as a time-lapse experiment.

The real achievement is that since its conception in early June, it has been stable enough to run in the background, capturing footage for such an occasion.

So here I present my latest time-lapse, a split screen video on the difference between a June day and a December day:

With time a solution also brings some observations and learning opportunities that would have been hard to imagine on conception:

Storage Capacity: In hindsight it was wise to start this project in June and during the summer solstice of the northern hemisphere. The maximised daylight hours created the most capture data stored to disk; therefore, the retention times of footage could be set to ensure the disk does not reach capacity.

While this works for MotionEye’s capture settings, the time-lapse script has been able to run without capacity constraints, so the 83GB (on a 3TB disk) will need to be moved / deleted off the disk as the nights grow longer to avoid exhaustion of storage space.

Camera Movement: During its conception, an elegant solution to create a makeshift mount for the camera was to fashion some bell wire (single core insulated wire) from the mount points of the Pi to the mount points of the camera, twisted in-between to add rigidity.

What I didn’t account for was the effect heat would have on twisted wire, this was apparent in the summer but seemed organic as the movement was in line with the temperature fluctuations of a normal day. However unforeseen was the winter and with that the part my central heating and radiators would manipulate the cameras position.

The combination of the radiator below the camera and the heating set on a timer resulted in large skews of the camera’s position, to the point where you could tell when the heating was on and off during the day. As a result, the footage used for winter was recorded while I wasn’t home and so the heating was off.