Using the railway network as a flatbed scanner

The article discusses a method of capturing images using a railway network as a flatbed scanner. This is achieved by pointing a camera out of a moving vehicle and capturing a single vertical line repeatedly, similar to the grayscale examples shown. By quickly stitching these lines together, a complete-looking image can be produced.

In the past, digital camera sensor technology was not advanced enough for this method, so digital scanning backs were developed in the 1990s. These devices capture high-resolution images without needing a large grid of pixels by moving a single line of pixels across the frame. Nowadays, image sensors are larger, but this approach remains cheaper for large formats than using a giant sensor.

The author initially wanted to build a digital scanning back for their large format camera but found building it too challenging. Instead, they decided to experiment with moving the camera instead and loaded film into their large format camera on top of a mountain in Vermont. They also captured a sofa using a phone on an office chair and wrote code to grab the leftmost column of each frame to create an image.

However, they soon realized that variations in speed would significantly affect the results. For their next attempt, they used the MBTA orange line and taped their phone to the seat to use its accelerometer, while holding their current phone to the window. They found that the accelerometer data was not useful, but the video still yielded interesting results.

The author also discovered a camera designed to read out its 1x2048 pixel image sensor at 19,000 times per second, which they obtained at a fraction of its original price. They used this camera to capture images on the railway network, utilizing its gigabit ethernet link and software.