General methods for the characterization of the nonuniformity of a radiometric calibration source

The spatial nonuniformity of a radiometric calibration source is key to its performance, uncertainty budget, and utility. Multiple disparate approaches are currently used to assess nonuniformity; thus, it is critical to understand how nonuniformity is calculated when comparing radiometric calibration sources. Two factors are key in assessing nonuniformity: (1) the methodology chosen to calculate the nonuniformity and (2) how the intensity of the light is sampled. Here, we present and compare different methodologies of calculating spatial nonuniformity. In general, we find that these methods can be categorized into two groups: those in which the equations assume a Gaussian population distribution and those in which the equations can be applied to any shape of distribution. Additionally, we compare different methods of sampling the light intensity within the region of interest. We show that peak deviation nonuniformity analysis, in which the pixels of the image of the illumination source are resampled into 69 equally sized bins encompassing the region of interest, can be used to effectively assess the nonuniformity of any radiance distribution. Furthermore, the nonuniformity determined by this method allows for an accurate comparison to industry standard nonuniformity metrics.