Optimization of data acquisition algorithm for temperature and emissivity distribution measurement using snapshot hyperspectral imaging systems

Abstract

Many industrial and scientific applications require precise measurement of temperature distribution over the surface of the inspected objects with a priori unknown emissivity. In these cases, imaging spectroscopy is an effective approach for quantitative thermography. This method is based on the measurement of thermal radiation spectrum in each pixel and its approximation of Planck's law by selecting proper values of temperature and emissivity. Snapshot hyperspectral imaging systems are especially interesting for this task as they provide an arbitrary number, positions and order of spectral bands and, therefore, enable multiple data acquisition modes. In this paper, we show that conventional equidistant location of spectral bands is not optimal in terms of temperature measurement error minimization. More effective approach is aggregation of the spectral bands into two narrow ranges located around particular wavelengths. We prove these statements by the theoretical consideration, mathematical modeling and experimental study on the tungsten plates using acousto-optical hyperspectral imager. Obtained results may be useful for the development and improvement of quantitative thermography systems with optimized data acquisition modes for various applications.

Keywords

• Thermography
• Snapshot hyperspectral imaging
• Mathematical modeling
• Approximation
• Measurement error minimization