This paper presents a UAV-based workflow for the inspection of buildings in cold-region climates, focusing specifically on rooftop snow depth estimation and thermal reconstruction of building envelopes. Accurate measurement of rooftop snow depth is critical, as an excessive snow load can lead to structural damage or roof collapse, resulting in substantial economic losses and potential fatalities. Thermal reconstruction can indicate thermal bridging in the building envelope, a phenomenon that can significantly reduce heating efficiency, causing increased energy consumption and higher utility expenses. Case studies conducted at multiple locations demonstrate the efficacy of UAV photogrammetry in accurately measuring rooftop snow depth, validated using manual field measurements and LiDAR scanning. We also propose a method of joint RGB-thermal reconstruction, capable of producing models with a high degree of geometric and radiometric accuracy without the requirement of homography or 3D transformations for image pair alignment. Results validate the approach through field measurements and pixel-wise thermal model evaluation. The proposed methods provide efficient, accurate, and safe alternatives to traditional inspection practices.