Intelligent condition recognition of transmission line based on digital image processing

According to the remote monitoring requirements of smart grid construction, in this paper the condition recognition approach combining digital image processing with artificial neural networks is proposed for transmission lines. The digital image processing methods, including gray scale transformation, histogram modification, wavelet packet denoising and edge detection are used to process the images of transmission lines and make the characteristics more outstanding. After dividing the images into some regions the distribution of edge features of transmission line components is extracted as characteristic values. This method has good adaptability. At last, a three-layer back propagation (BP) artificial neural network (ANN) is constructed and applied recognize the typical transmission line conditions. The result shows that this approach has good recognition rate and popularization.