Power infrared image segmentation model integrating C-V and LAC active contours

Infrared image segmentation of power equipment is one of the key technologies for power equipment status monitoring and fault diagnosis. However, due to the presence of complex backgrounds, noise interference, and diverse target shapes in the images, traditional methods are difficult to achieve ideal segmentation results. To address this issue, this study proposed a power image segmentation model that integrated C-V (Chan Vese) and LAC (localized active contour) active contours based on the active contour segmentation model. This model introduced p-Dirichlet regularization, integrating the LAC local model and C-V model into a horizontal set to construct the final contour evolution equation. The simulation experiment results show that the average intersection over union (IoU) of the paper model reaches 0.933, and the average Dice similarity coefficient is 93.4%. The model exhibits excellent performance in power equipment image segmentation tasks, achieving effective segmentation of grayscale non-uniform images and complex background power images.