Building Height Prediction with Instance Segmentation

TL;DR

This paper introduces a novel method for predicting building heights from single RGB satellite images using instance segmentation, achieving promising accuracy without relying on DSM data.

Contribution

The study presents a new instance segmentation approach for building height estimation from RGB images, utilizing transfer learning and achieving competitive accuracy.

Findings

Bounding box mAP of 59

Mask mAP of 52.6

70% accuracy in height classification

Abstract

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.