Inspiration
Our team consists of 3rd year physics bachelour students: Gusts, Rauls and Rūdolfs. We wanted to apply the knowledge we have gained during our physics studies to this project. There was a brainstorming session about ultrasound for mine detection, but determined it too impractical. After that, we figured something in navigation could be our project - the Sun seemed managable :D
What it does
The program outputs approximate geographical coordinates of an observer by taking two Sun altitude measurements at specific times using special algorithms.
Extra details- along side taking the pictures, you also capture the time when the photo was taken and the accelerometer values when the camera has the Sun at its center (for both pictures). With this you can pinpoint your location to two points.
How we built it
We did a proof of concept- we took the two pictures of the Sun at our Hackathon building and tried to determine the coordinates of this building.
Challenges we ran into
Big errors in coordinates, especially geographical longitude (for our pictures). Also could be a lack of organization. At some points we were doing the same thing... There were points where we all were stomped on what to do and jokingly went to the toilet/for a walk frequently.
Accomplishments that we're proud of
We found a geometrically compelling way of deriving the location of the photographer, if two pictures are taken. Using online available data, we got 15 km away from our location!
We endured this challenge and did not give up. We spent quality time together, successfully got annoyed into focus and creativity.
What we learned
Sad truths about astronomy and that it is not as simple as we thought... Like how the day/night line is not a straight line in the usual Mercator map projection, where lattitude/longitude angles are 90 degrees. It is only a straight line during equinoxes, when day and night are the same length.
After building multiple models with similar mathematical background, we concluded that basing the model on subsolar points (where Sun is overhead) was the most effective one. The concept of subsolar point is foreign for our region as Sun is never directly above us and only during this Hackaton we found out about this.
What's next for Sonavlar
The prelimenary results are promising, with best calculated results being as close as 15 kilometres from the observer. Next logical step would be improving the algorithm by taking into account known phenomena (e.g., Earth having eliptic orbit around the Sun instead of circular) and making a GUI.
However, if we were to continue, we would advance the prototype- try more measurements, get a better understanding of the error. We would like to make an app where the user could do this without being a nerd. :)
Log in or sign up for Devpost to join the conversation.