Aircraft Wake Detection Method Based on KNN and Morphology
In order to improve the intelligence of the air traffic management system,the problem of aircraft wake turbulence detection is solved.Combined with Doppler lidar technology,an aircraft wake turbulence detection method based on K-Nearest Neighbour(KNN)algorithm and image morphological processing technology is proposed.Firstly,the radar is used to scan the inbound and outbound airspace of airport to obtain the wind field data of the target region.Secondly,the dynamic sliding window with different window sizes is used to generate the proposed candidate regions for the radial velocity field of the acquired aircraft wake stream.Finally,the morphological features of the wake stream in the candidate regions(ROIs)are extracted into the K-Nearest Neighbour algorithm using the top-hat and black-hat morphology operations to detect them and are compared with those of the same conditions using the detection method based on the wake velocity extreme difference feature method.The experimental results show that the accuracy of the proposed aircraft wake detection method is 22.58%,9.29%and 14.22%higher than the accuracy,recall and F1-score of the speed polarimetric-based method,respectively,and the method can provide decision support for the controllers.
wake detectionmorphologyKNNtarget detectionDoppler lidarvisualization
万方数据
维普
