Infrared Image Edge Extraction of Cable Terminal Based on Improved Eight Direction Sobel Operator

The infrared thermal image is an importance in condition monitoring of electrical equipment. However, the infrared image edge is coarse, fuzzy, and with serious noise and brings great inconvenience to the infrared image processing. An adaptive optimal threshold edge extraction algorithm based on improved Sobel operator is proposed in this paper. In the algorithm, by eight-direction Sobel operators the infrared image edge is extracted in the high temperature area in noisy environment. At the same time, the wavelet coefficients in sub-band are described by general Gaussian distribution and the variance is estimated from the local neighborhood information of sub-band wavelet coefficients so that the adaptive optimal denoising threshold can be obtained. Finally, the edge extraction infrared image is denoised by combining with improved Sobel operator and the optimal threshold. Matlab simulation results show that the algorithm can effectively detect the edge of infrared image and greatly improve its anti-noise ability.