Cable pipeline fault discrimination algorithm based on FCN full convolutional neural network

The invention relates to a cable pipeline fault discrimination algorithm based on an FCN full convolutional neural network. By using the FCN full convolutional neural network algorithm, the fast batch identification operation of the pipeline contour can be realized. And automatically identifying the roundness of the cable pipeline according to the obtained contour shape, and distinguishing different fault types according to the shape difference so as to obtain the classification condition of the fault pipeline. According to the invention, the FCN full convolutional neural network algorithm is applied to the field of cable pipeline fault identification, fault identification and specific state classification of the pipeline can be realized, the cable pipeline fault can be rapidly and conveniently identified, and the overall efficiency is improved. 本发明涉及一种基于FCN全卷积神经网络的电缆管线故障判别算法。利用FCN全卷积神经网络算法，能够实现管线轮廓的快速批量识别操作。将得到的轮廓形状进行电缆管线圆度的自动辨识，根据形状差异区分不同的故障类型，进而得到故障管线的分类情况。本发明将FCN全卷积神经网络算法应用于电缆管线故障识别领域，能够实现管线的故障识别和具体状态分类，有