Research on Improved Algorithm of Infrared Image Edge Detection for Electrical Equipment

In order to diagnose the thermal anomaly of electrical equipment in time, an improved Canny edge detection algorithm for infrared image of electrical equipment is proposed to facilitate fault location. To improve the dexterity and adaptive ability of Canny edge detection algorithm, wavelet transform and improved homomorphism are used Instead of Gauss filter, the filtering algorithm can denoise the image, reduce the influence of noise and light, keep the edge intensity and details; improve Sobel operator, increase the gradient amplitude in 45
⁰
and 135
⁰
directions, so as to improve the edge positioning effect; use non maximum suppression to refine the image, and use Otsu algorithm according to the image gray level to adaptively obtain high and low thresholds, and use high and low thresholds Value algorithm detects and connects edges. Through the image verification of electrical equipment, the experimental data shows that the improved algorithm has better detection accuracy and stronger adaptability than Canny edge detection algorithm.