contrast enhancement

In thermal barrier coatings (TBC), surface cracks, debonding, and thickness degradation may occur during the manufacturing process or life cycle, leading to poor performance and ultimately a dangerous system failure. The main goal of non-destructive testing of thermal barrier coatings is to detect these defects and determine the health of the coating. Various non-destructive inspection methods have been proposed to evaluate thermal barrier coatings, and due to the numerous advantages of thermography, including high speed, low cost, safety, no need for direct contact, automation capability, and inspection of a large area of ​​the part, this method has received special attention from researchers. This study will present a method for manufacturing samples with different diameters of artificial separation defects. The following is the equipment's arrangement and the sample's thermography process. It was concluded that blackening the surface of the sample by increasing the amount of thermal energy absorption increased the ability to identify separation defects and increased the signal-to-noise ratio by 257%. Finally, by implementing different filters on the recorded raw thermal images, it has been shown that in both cases the best filter in terms of SNR is the median filter and then the Gaussian filter. The background removal filter also had no noticeable effect on increasing the signal-to-noise ratio and acted as a complement to the median and Gaussian filters by reducing the fixed error.

In this article, a contrast enhancement algorithm using wavelet transform is proposed to create natural enhancement in images with various brightness ranges. In the proposed method, the input image is first decomposed into 4 sub-bands by discrete wavelet transformation. Then, in the LL sub band image, the histogram cutting process is performed by the cutting threshold level equal to the average intensities. Next, the cut histogram based on entropy is divided into three parts with approximately equal number of pixels, and before equalization process, each sub-histogram is mapped to the new dynamic range. Finally, we use the inverse discrete wavelet transform to create an improved image. By controlling the enhancement ratio, the proposed method produces an image with maximum details and natural enhancement in the output image. Comparing the performance of the proposed method with previously presented methods, in terms of entropy as well as visual quality based on the mean opinion score, shows the superiority of the proposed algorithm.