FOD Detection Algorithm Based on WGMF and SVDD

Using millimeter wave radar to detect foreign object debris (FOD) in airport runway is a popular solution in the field of civil aviation safety. Due to the existence of non-uniform clutter, the traditional detection method based on constant false alarm rate has the problems of high false alarm rate and poor detection performance. A weighted generalized matched filtering (WGMF) combined with support vector data description (SVDD) method is proposed. Firstly, WGFM is used to preprocess the radar recorded FOD echo data to suppress clutter and reduce the false alarm probability. Then, bispectral features are extracted from the echo data to realize the transformation of the data from the original echo domain to the feature domain. Finally, FOD detection is realized by SVDD classifier. The experimental results show that the proposed method can effectively reduce the false alarm probability and improve the detection performance.