In a unique project, a New York Engineer has introduced dropofahat.zone, a system leveraging artificial intelligence to toss baseball caps to individuals passing by his apartment window.
The concept allows anyone to select a preferred time slot, purchase a baseball cap, and position themselves beneath the engineer’s window for three seconds to catch a falling cap. The caps are released every five minutes, with a video demonstration showcasing the system in action.
Utilizing a setup consisting of RASPBERRY PI, ADAFRIIT step engine, and online tool marking, the engineer opted for lightweight caps equipped with propellers. The resetting mechanism relies on yarn wound around the step engine for functionality.
Upon detection of a person below the window, the AI triggers a specialized Python script on the Raspberry Pi, identifying the presence of a passerby standing underneath.
The AI training process involved capturing a few minutes of video footage from the window to teach the system to detect pedestrians without hats. The engineer shared a collection of training data comprising 133 annotated images and a model combination.
The “Dropofahat.zone” initiative marks the initial phase towards realizing a larger vision of transforming building windows into virtual storefronts, capable of dispensing various items to pedestrians at any hour of the day or night.