Building Height Prediction with Instance Segmentation
This work addresses the challenge of extracting building heights for applications like telecommunications and city planning, but it is incremental as it builds on existing instance segmentation techniques with transfer learning.
The study tackled building height prediction from single RGB satellite images by developing an instance segmentation-based method, achieving a bounding box mAP of 59, mask mAP of 52.6, and 70% average accuracy for height classes in the test set.
Extracting building heights from satellite images is an active research area used in many fields such as telecommunications, city planning, etc. Many studies utilize DSM (Digital Surface Models) generated with lidars or stereo images for this purpose. Predicting the height of the buildings using only RGB images is challenging due to the insufficient amount of data, low data quality, variations of building types, different angles of light and shadow, etc. In this study, we present an instance segmentation-based building height extraction method to predict building masks with their respective heights from a single RGB satellite image. We used satellite images with building height annotations of certain cities along with an open-source satellite dataset with the transfer learning approach. We reached, the bounding box mAP 59, the mask mAP 52.6, and the average accuracy value of 70% for buildings belonging to each height class in our test set.