Abstract Urban thermography is a non‐invasive measurement technique commonly used for building diagnosis and energy efficiency evaluation. The physical interpretation of thermal images is a challenging task because they do not necessarily depict the real temperature of the surfaces, but one estimated from the measured incoming radiation. In this sense, the computational rendering of a thermal image can be useful to understand the results captured in a measurement campaign. The computer graphics community has proposed techniques for light rendering that are used for its thermal counterpart. In this work, a physically based simulation methodology based on a combination of the finite element method (FEM) and ray tracing is presented. The proposed methods were tested using a highly detailed urban geometry. Directional emissivity models, glossy reflectivity functions and importance sampling were used to render thermal images. The simulation results were compared with a set of measured thermograms, showing good agreement between them.

Collaborative work is very important in engineering and architecture because of the many disciplines involved in the design process. For this reason, it is necessary to deliver to all the actors of the system not only relevant data, but also information corresponding to their specific skills. Because in architecture the visual experience is mainly based on projections, we want to show that a wider choice of them is the best way to improve the data exchanges. We choose the simple geometric problem of solar illumination. Because it is essential to take it into account from the first steps of the design to the final project, the partners have to communicate with several kinds of projections that are not always the most standard ones. In this process all the collaborators are expected to contribute to the fluency of the exchanges with the consequence that, in order to magnify some aspects of the design, they are induced to define new representations.