Edition Sept 21st-23rd, 2018
Map the Alps
Map the alps in two days. We approached to rapidly utilize image processing and machine learning to achieve our goals.
"Our goal was to create a Tensor flow data model in order to support Techno Alpin in planning of snowmaking systems in the most efficient way.
Such module will be helpful in creation of ski area maps and equipment.
Modern approaches and tools aid at retrieving information from images. Therefore the motivation was to develop an approach and use it to solve the problem statement.
How did you build it?
We used open source libraries(OpenCV, Anaconda, TensorFlow, Google Earth Pro) to chain them into a powerful pipeline.
short time frame, high complexity & missing training data sets.
Unfortunately, the training and implementation of the correct learning function was not possible in the given time frame. We achieved the creation of a module which could identify pictures of artificial lakes (within our collected sample/test and training data).
For further development a more theoretical approach would be good for the start but then we could create good models that can identify some of the objects where focused by the challenge."