Projects

RGB Point Cloud and Thermal Information Data Fusion for Building Thermal Map Modeling

  • Reconstruct 3D models using high-resolution RGB images and project thermal information onto the 3D models.

  • Investigate how the RGB and thermal data fusion approach performs under various conditions when modeling large areas.

  • Comparison between data fusion performances of tie points and all points in a 3D model.

  • Provide strategies for improving the accuracy of the data fusion approach in terms of UAS flight configurations.

figure1.png

Building Object and Outdoor Scene Segmentation (BOOSS) - Multi-channel (RGB + Thermal) Aerial Imagery Datasets

  • We proposed to detect thermal bridges on roofs.

  • We proposed to study differentiating salient components, such as facades and roofs where energy loss was important to monitor, as well as from peripheral components such as cars and equipment where heat loss was not our focus, but which may interface with the study.

  • Please visit 

  • https://zenodo.org/record/5241286#.YWeazhrMKiN

Semantic Point Segmentation

  • Our project aims to provide an extensive database of annotated ground truth point clouds reconstructed by aerial image photogrammetry.

  • Training and validating 3D segmentation.

  • https://www.stpls3d.com/

Untitled.png