Partial Differential Equation Based Enhancement And Segmentation Method For Infrared Images Of Electrical Equipments

In the power systems, most of electrical equipments work in harsh environmentswhich usually cause accidents. Which not only results in damaging equipments,interfering normal production, losing the economic interest of electrical enterprises,but also influences the safe operation of electric systems that endanger the lives oftechnicians. To ensure that the system is running in security, it is necessary to find outthe equipment which has already or trends to malfunction.This paper takes advantage of partial differential equation to enhance andsegment the collected infrared images. For denoising infrared images while retainingthe detail of image edge, the diffusion function of classical P-M model is improved.In the improved P-M model, when the gradient magnitude of image is within therange of wavelet threshold, the improved diffusion function is the weighted sum ofthe classical diffusion function and exponential function. The experiment of presentpaper indicates that the improved model is superior to classical P-M model, Medianfilter and Wiener filter in image denoising and edge retaining. This method adopts thepartial differential equation to enhance the contrast of histogram equalization modeland it solves the problem that noises in infrared images are greatly raised while usingthe traditional histogram equalization in contrast enhancement. The experiment alsoindicates that the image contrast of infrared images from electrical equipment islargely improved by using the improved method. This method adopts the improvedGAC model to segment the infrared images and its segment process is actually theevolution process of a closed curve. Compared with the traditional segment method,the improved GAC model has no problem of image edge fracture while segmentingimages. And the experiment indicates that the improved GAC model solves theproblem that the GAC model always remains in the local minimum while segmentingimages which lead to unsatisfactory or unsuccessful segmentation.In addition, the finite difference method is used to discretize the model while solving the partial differential equation. Transforming the partial differentialequations into the iterative equations greatly reduces the calculation and shortens theexecuting time of programs. It realizes the enhancement and segmentation of infraredimages from electric equipments, provides guarantee for the application of theinfrared thermography technology in intelligrid.