Fairy Caves

I started out with a kernel of an idea, a seed that I hoped would grow into something more. I had researched and planned but I didn’t quite know where my process would take me. I started to experiment, I took what I knew and dug deeper, searching for some kind of spark. I created a system that tracks movement in real time. It was a bare bones hack riddled with inconsistency, but it had a hint of something more. This is the moment I realized what I needed to create. I centered my focus on one singular question: what is the difference between technology and magic. This world is the result.

In the beginning, I didn’t know what kind of interaction I wanted to create. I knew I was interested in applying the capabilities of open source computer vision technology, so I began to experiment with was possible. During the process of fine-tuning a motion detection system, I had a realization: without the insight into the processes in the code, the result felt almost magical. This little box seemed to have a mind of its own. With this in mind, I began to further obscure the processes of the device. This intentional obfuscation gradually increased the sensation of other-worldliness. Installing this minimalist system in a darkened room with a bit of preamble text worked to instill a curious state of mind in the users.

The systems employed by Fairy Caves are a straightforward python script utilizing OpenCV algorithms deployed on a Raspberry Pi. A Pi Camera Module is used so that the image exposure can be fine-tuned to minimize post-processing. The LED torches play a critical role in this process. Because the LED appears as a point source, the light “pushes” through the underexposed image to providing a high contrast image suitable for the tracking algorithms.

The RasPi was chosen over more powerful computers because of its small size and the camera's versatility and customization options. The system was intended to be capable of being deployed in a variety of environments without the need for long cable runs or custom installation hardware. The processing power of the Pi was stressed throughout the process and limited the quality of graphics that could be output. While this limit was unfortunate, I don’t think It significantly impacted the resulting interaction. A future install could utilize a separate, more conventional, projection mapping system to increase the quality of the outputted graphics.

I thoroughly enjoyed this process and I appreciated the willingness of my professors to allow me to engage in a more experimental design. In addition to the advancement of my story building skills, I was able to develop skills in employing OpenCV technology and Python in general. I think that the process of exploring technology first led to a more compelling final experience. Without this initial exploration phase, I would not have known the possible applications of the technology in cultivating an experience. While I don’t want to claim that the final product was implemented to an infallible level, I am satisfied with the robustness I achieved with limited time and resources available. The feedback I received from peers and instructors was mostly in-line with my own personal knowledge of the imperfections in this project. I was surprised with the extent and duration that classmates engaged in the work. I was surprised by the number of people that attempted to push the limits of the project, and even “break” the technology. I think this is an extremely useful insight to keep in mind for future work. Overall, I feel that I was successful in this project and gained knowledge in a variety of areas from both the technology and storytelling facets of the work.