Chirplet parameter estimator based on ellipse fitting in time-frequency distributions for ultrasonic NDE applications

Estimating the parameters of chirp signals is essential in several important applications such as radar, sonar and ultrasonic imaging. Chirp pulses characterize dispersive media, and they are also very effective for pulse compression and improved echo detection. Furthermore, in Doppler ultrasound velocimetry, frequency-change associated with moving object can be characterized by chirp rate estimation. Time-Frequency Distributions (TFDs) of signals with chirplet components result in elliptical contours. In this study, a chirplet parameter estimator by means of ellipse fitting in time-frequency plane is developed. This estimator performs robustly in presence of the noise and offer much lower computation complexity compared to conventional methods.