Aircraft Wake Recognition Based on Improved ParNet Convolutional Neural Network

The occurrence of wake can pose a threat to the flight safety aircraft and affect runway capacity airport operation efficiency. To effectively identify wake, this paper proposes novel convolutional neural network (CNN) method recognition based on improved parallel (ParNet). Depthwise separable convolution (DSC) was introduced into ParNet make model lightweight. In addition, block attention module (CBAM) improve extract spatial features wind field. proposed used lidar field scanning image Hong Kong International Airport. best effect obtained with accuracy 98.91% an F1 value 98.90%. As number parameters only 0.46 M, could be identified ordinary computer. Thus, wake.