Automatic Construction of Probabilistic Dynamic Bidirectional Reflectance Distribution Functions from Reflection Screen Images

摘要

The reflections of high energy laser off surfaces can present hazards to persons and instruments at significant distances. The heating from these lasers cause changes in the reflection characteristics of surfaces they impact. As such, the reflections from these surfaces cannot be properly modeled with static bidirectional reflectance distribution functions (BRDFs), but require time-dynamic BRDFs. Moreover, the time-evolution of the surface reflections is not deterministic, but can vary even when the materials and irradiance conditions are nearly identical, such that only probabilistic characterization is realistic. Due to the swiftly changing nature of the reflections, traditional BRDF measurements with goniometric instruments is impossible, and BRDFs must be deduced from images of the reflected light incident on a screen which intercepts a portion of the reflection solid angle. A model has been constructed to describe these complex probabilistic dynamic BRDFs with only a moderate number of intuitive parameters, where these parameters have central values and statistical variances. These simple parametric representations are appropriate for use in predictive modeling codes and are also easily adjustable to allow facile exploration of the sensitivity of hazards to laser, material, and model uncertainties. An automated procedure has been created for determining appropriate parameter values and variances from captured screen images, without the need for case-by-case human judgment. Examples of the parameter determination procedure are presented.