Segmented fast linear canonical transform

Investigation of the discrete and fast linear canonical transforms is becoming one of the hottest research topics in modern signal processing and optics. How to handle and obtain the linear canonical frequency spectrum of very large input data based on equipment with limited memory space is one of the key problems. To focus on this problem, a new kind of segmented fast linear canonical transform has been proposed in this paper. First, the large data is segmented into short data. Thereby, the proposed algorithms can calculate very large input data and simultaneously keep the ideal frequency resolution. Second, the complexity of the derived algorithms has been analyzed in detail for different kinds of signals. Their performance with regard to resolution and precision are compared with the existing fast linear canonical transforms. Finally, experimental results are presented to verify the correctness of the results obtained.