Contrast Enhancement Method of SF₆ Infrared Image Based on Tri-histogram Equalization Algorithm

Infrared images of sulfur fluoride (SF₆) leaked gas in power equipment are easily merged with the background, and it is difficult for humans to identify the leaked gas in low-contrast images, rendering it difficult to maintain power equipment. A low-contrast enhancement method for SF₆ leaked infrared image based on tri-histogram equalization is proposed. First, a cubic spline interpolation is used to fit the image histogram to construct a second-order continuous curve, and the absolute value of the first derivative of the curve corresponding to each gray level is calculated. Second, the two extreme points of the histogram are calculated according to the absolute value distribution of the first derivative and histogram. The extreme points divided the histogram into two peaks and relatively flat valley regions. Finally, the image is divided into three sub-images according to the tri-histogram, and the three sub-images are histogram-equalized. The sub-images are merged into a new image, which is the enhanced image. To verify the validity of our algorithm, an infrared video of power equipment with SF₆ leakage, which is recorded in substation fields, is tested to enhance the contrast. Our algorithm is compared with the CLAHE and bi-histogram equalization algorithms. The experimental results show that our method enhances not only the globe contrast for infrared images but also the local contrast for leaking gas. Therefore, the infrared image of the SF₆ visual effect is improved. Compared with CLAHE and the bi-histogram equalization method, our method obtained a higher peak signal-to-noise ratio and root-mean-square contrast values, and the enhanced image quality of our method was better than that of the other methods.