3D shape from thermal patterns: investigation of projection parameters in simulation and experiment

Structured light projection techniques based on diffuse reflection are widely used for accurate, fast, contactless, and nondestructive optical 3D shape measurements. It cannot be utilized to measure uncooperative materials, i.e., materials with optical properties such as being glossy, transparent, absorbent, or translucent. Recently, it was shown that 3D reconstruction of an uncooperative object can be performed by a two-step process for each camera image pair. In the first step, the object absorbs a projected thermal pattern, e.g., in the long-wave infrared range. In the second step, after energy conversion, the object surface reemits light according to Planck's law. Whereas the irradiation can be performed by a CO2 laser at 10.6 mu m, the detection of the reemitted light can be carried out by mid-wave infrared (MWIR) cameras sensitive in the wavelength range between 3 and 5 mu m. In order to achieve accurate 3D results in a short measurement time, the projection parameters like radiation intensity and illumination time as well as the projection patterns have to be optimized depending on optical and thermal material properties (e.g., complex spectral refractive index, thermal conductivity, specific heat capacity, emissivity). Therefore, we have developed a simulation tool based on the Beer-Lambert law (for absorption of the irradiation) and on the heat diffusion equation (for the illumination-induced thermal pattern on the object surface). In this contribution, we present our simulation tool and several simulation results. We apply our tool to investigate the projection parameters and projection patterns for a given material and a specific total measurement time. Finally, we experimentally verify the theoretical results with our MWIR 3D sensor.