Pipeline miter joint angle detection method and device

The invention provides a pipeline miter angle detection method and device, and the method comprises the steps: collecting magnetic flux leakage curve samples of different miter angles in a pipeline through a magnetic flux leakage inner detector, and dividing the magnetic flux leakage curve samples into a training image sample set and a test image sample set; inputting the training image sample set into a convolutional neural network to carry out feature recognition learning training so as to construct an initial convolutional neural network model; calibrating the initial convolutional neural network model by using the test image sample set to obtain a final convolutional neural network model; and acquiring a magnetic flux leakage curve image of the miter joint angle of the to-be-detected pipeline and inputting the image into the final convolutional neural network model to obtain the range of the miter joint angle of the to-be-detected pipeline. According to the method, the image recognition advantage of the g