GeBDA: Building Damage Assessment as Text-Based Sequence Prediction
2026-08-28 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors explore whether a general Vision-Language Model (VLM) can identify buildings and assess their damage just by generating sequences of text, instead of using specialized systems. They treat damage assessment as predicting a list of boxes around buildings, each with coordinates and a damage label. Using the Gemma model and pairs of satellite images taken before and after an event, their initial results show the model can map damage reasonably well with the right text instructions.
Building Damage AssessmentVision-Language ModelAutoregressive Sequence GenerationBounding BoxesBi-temporal Satellite ImagesGemma ModelDamage MappingFine-tuningGeospatial Imaging
Authors
Olivier Dietrich, Krishna Sapkota, Konrad Schindler, Genady Beryozkin
Abstract
Conventionally, Building Damage Assessment (BDA) is tackled either with dedicated network architectures or by fine-tuning geospatial image foundation models. In this work, we ask whether a general-purpose Vision-Language Model (VLM) can localize buildings and grade their damage through autoregressive sequence generation alone. We cast BDA as predicting a variable-length set of bounding boxes, each specified by its coordinates and a damage label. Our preliminary implementation, based on the open Gemma model, achieves promising damage mapping results from only bi-temporal satellite images and a suitable text prompt.