Arccos is the leading provider of advanced analytics and machine learning insights for the golf industry. Watch how Sal Syed (CEO, Arccos) and his team partnered with Microsoft to use machine learning to help golfers make better decisions on the course.
The goal was to enable Arccos users to make real time shot decisions when faced with limited playable areas on the course as well as obstacles such as trees, hazards, and other restrictions. The project was divided into three machine learning use cases: cloud detection, playable area labeling, and semantic image segmentation.
The team used Machine Learning in Azure to screen Apple Map GPS images to detect obstructions caused by cloud cover and automatically substitute affected images with a version from Google Maps. The model was implemented in Python using the Cognitive Services Custom Vision Service. Next, the team analyzed satellite images of golf courses to detect playable areas by distinguishing the fairway and green from the rest of the area, e.g. trees, shrubs, buildings, etc. Otsu thresholding and watershed algorithms were implemented to create a ‘playable area’ data set with pre generated labels to train the machine learning model. The final part of the project was semantic image segmentation for tree detection utilizing Keras deep learning framework and a U-Net model.
Key Technologies Used: Cognitive Services Custom Vision Service, Azure Notebooks, Azure Deep Learning VM, Python, Keras, Scikit-Learn
Get Started with CTNK: https://aka.ms/getstartedCNTK
Understanding Image Segmentation: https://aka.ms/ImageSegmentation
Subscribe to Microsoft on YouTube here: https://aka.ms/SubscribeToYouTube
Follow us on social:
LinkedIn: https://www.linkedin.com/company/microsoft/
Twitter: https://twitter.com/Microsoft
Facebook: https://www.facebook.com/Microsoft/
Instagram: https://www.instagram.com/microsoft/
For more about Microsoft, our technology, and our mission, visit https://aka.ms/microsoftstories
