Infrared Image Segmentation Method Of Substation High Voltage Equipment

The infrared image of electrical equipment has low resolution and uneven gray change, and can be easily affected by noise. Considering the disadvantage of infrared image and advantage of classical random walker algorithm with high computational efficiency and closed segmentation boundary, this paper designs two improved methods: automatic random walker based on IFS(Intuitionistic Fuzzy Sets) edge and an infrared image segmentation method based on L0 smoothing and random walker, and realizes the application of the algorithm of Lazy Random Walker(LRW) in the field of electrical equipment infrared image segmentation.The method of automatic random walker based on IFS edge detects the image edge information by IFS edge extraction method and acquires the connected domains information of image edge by morphological method. It selects the centre point of connected domains as the seeds of random walker and segments the edge image by random walker to obtain the target location. In order to verify the effectiveness of the proposed method, this paper sets up multiple sets of contrast experiment. The experimental results show that this method has realized automatic random walk image segmentation, and improved the accuracy of image segmentation in some degree.The infrared image segmentation method based on L0 smoothing preprocesses the infrared image by using the smoothing method based on L0 gradient minimization, and then segments by random walker. The results show that the smoothing method based on L0 gradient minimization can smooth the image internal detail in the case of the prominent edge and reduce the influence of local image gray value change to the segmentation path of random walker. In the experiment, the proposed method performs very well on removing noise and greatly improves the accuracy of the classical random walker on the substation equipment infrared image segmentation.Finally, the paper studies the LRW algorithm in the application of infrared image segmentation of electrical equipment, and completes infrared image segmentation experiments under different conditions and electrical equipment. The results verify that the LRW segmentation algorithm has higher robustness compared with the classical random walk algorithm.