Spatial-Data-Science-and-GEO-AI-Lab/Park-NET-identifying-Urban-parks-using-multi-source-spatial-data-and-Geo-AI

Project to investigate and develop spatial data-driven Geo-AI models (Convolutional Neural Network) to identify urban greenspace from Satellite images by integrating multiple data sources (e.g. vector data of urban parks)

Park-NET: identifying Urban parks using multi source spatial data and Geo-AI

This is a master thesis project done by Marta Kozłowska and Jiawei Zhao for ‘Applied Data Science programme at Utrecht University. We have a goal of analysing to what extent can a reproducible CNN model that predicts urban greenspace based on open source be created. The process involved creating two CNN models. One was the U-Net model built from scratch, and the other was the U-Net with ResNet34/50 encoder to make use of the transfer learning approach.

Go to our separate folder to see which methods exactly we used and what results we got.