Evaluation of the temperature pattern of a complex body from thermal imaging and 3D information: A method and its MATLAB implementation

The standard setting of a camera used in Infrared thermography (IRT) is based on the choice of the same values
of emissivity and distance for all pixels of a thermal image even if the emissivity depends on the relative position
of camera and observed surface. Often this is not a problem. However, the resulting temperature pattern could be
inadequate if a body having a complex shape is observed from strongly constrained positions. In order to face
this issue, a procedure aimed at providing a correct temperature pattern by using 3D information related to a
point cloud is proposed together with its MATLAB implementation (COMAP3 toolbox). For each pixel of a
thermal image, the relative position of camera and observed surface is estimated, leading to pixel-specific values
of emissivity and distance. The temperature obtained in this way is also mapped onto the point cloud. The
effectiveness of the procedure in recognizing areas characterized by peculiar thermal behavior is shown in the
case of a historic cylindrical masonry bell tower (Caorle’s bell tower, Venice, Italy).