Image processing based on the detection of external defects of fan tower Weld

With the continuous development of wind power generation technology and increase in demand for electric energy, height fan tower is more demanding. It very important to detect weld produced welding process tower. In this paper, an algorithm defect detection based on traditional image processing convolutional neural network proposed. Firstly, used gray collected by industrial camera. Then, seam enhanced improve visual effect clear image, which convenient further analysis computer. Finally, as input trained convolution judge whether there are defects outside weld.