Classification for golf ball by 'Impact-sound' features based on wavelet spectrum

摘要

We studied a cool sound a golf ball makes after being struck and the relation between the impact sounds and physical properties of an object for developing golf balls. We proposed a feature extraction method for these impact sounds based on a Fourier transform in our previous research in order to analyze the impact sounds emanating from struck golf balls. However, the method based on a Fourier transform does not have a sufficient enough time resolution to effectively analyze impact sounds because the energy of these sounds is a short and quick. We propose a feature extraction method based on the wavelet transform to overcome this problem. We carried out an analysis of the impact sounds from a golf ball in order to investigate the relation between the impact sounds and the physical properties of golf balls. We recorded ten kinds of golf ball impact sounds using four kinds of golf clubs (a driver, a putter, a six-iron, and a sand wedge). Ten balls were classified into three categories with a k-means by using the features that were extracted from the impact sounds. We investigated the relation between these sounds and the physical properties of the golf balls using the classified results. The results from the investigation indicated the correlations between the cover hardness and the impact sounds of the driver and putter.