Perceptual hash-based dimensionality reduction method, recognition method and recognition system for cyclostationary signals

• Main Authors: DOU RENJIE, MEN XIUHUA, ZHANG HUI, LIU HAINING, LIU CHENGLIANG
• Format: Patent
• Language: Chinese; English
• Published: 26.10.2018
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a perceptual hash-based dimensionality reduction method, recognition method and recognition system for cyclostationary signals. The dimensionality reduction method includes thefollowing steps: extracting primary functions from known cyclostationary signals of each device state to form a device state dictionary based on a dictionary learning algorithm, merging the primary functions of all device states to obtain a device state redundancy dictionary; performing sparse decomposition on to-be-detected cyclostationary signals by utilizing the device state redundancy dictionary based on a sparse coding method to obtain sparse representation of the to-be-detected cyclostationary signals; calculating activation feature vectors of the primary functions in the device state redundancy dictionary based on the sparse representation of the to-be-detected cyclostationary signals; converting the activation feature vectors of the primary functions into a binary sequence; and converting the obtained bin