Analysis of transmitted ultrasound signals through apples at different storage times using the continuous wavelet transformation

Abstract

We investigate the transmitted ultrasonic signals through apples at different storage times using continuous wavelet transform. Apples are comprised of the flesh, ovary and rind, and the materials alter the signal travelling through them. The continuous wavelet transform enables us to visualize temporal frequency components, and to investigate the variation of signals due to the dispersive wave propagation. We model time dependent transitions of low frequency with a high intensity by using an exponential function which has three parameters, and these parameters are estimated by a least square method for the signals obtained at different storage times. The regression results show that a high correlation exists between frequencies and storage time. Moreover, the Dunnett's t-test results of regression parameters show that there are significant variations in two regression parameters after 15 days, and the last parameter deviates only after 23 days later. These results indicate that the continuous wavelet transform of the transmitted ultrasonic signals can be used to monitor the changes of mechanical properties of apples, such as softening, during the storage time.