Image processing for frost thickness measurement in fin-and-tube heat exchangers

In studies addressing performance degradations owing to frost formation on heat exchangers, the frost thickness is considered an important parameter. The most commonly used method for measuring frost thickness comprises capturing and analyzing an image. However, relatively few studies have investigated the most reasonable method for deriving frost thickness by analyzing an image. In current measurement techniques, it is difficult to quantitatively determine which methods have less error relative to the true value in a complex shape. Therefore, this study attempted to recommend a more reasonable method, based on analysis of various aspects. First, the fundamental concept of image processing was described in detail. Then, the effects of thresholding algorithms, threshold application algorithms, and the determination of the number of segments on the frost thickness were investigated. Based on the results, an appropriate image processing method for measuring frost thickness was recommended. The frost thickness obtained by this method was well-measured, with an uncertainty of approximately 0.18 mm.