The intersection between artificial intelligence (AI) and GIS is creating massive opportunities that weren't possible before. This video by Rohit Singh, Esri Inc showcases several examples of applying AI, machine learning, and deep learning to geospatial data using ArcGIS API for Python. It covers several scenarios of applying the latest machine learning and deep learning techniques to geospatial data, including the following:
- Detecting objects using satellite imagery such as locating swimming pools from satellite imagery using fast.ai and PyTorch, classifying healthy and neglected pools, and visualizing the results on web maps
- Detecting objects in videos using TensorFlow such as traffic volume and anomalies using live camera feeds
- 3D building shapes reconstructed from remote-sensing data
- Semantic segmentation (such as land-cover classification and identifying roads and building footprints) using satellite imagery and models such as U-Net and Mask R-CNN
- Enhancing imagery using super-resolution networks and increasing the clarity and zoom levels
- Generating maps from satellite imagery using CycleGAN
- Map art generation and cartographic style transfer using neural style transfer
