Authors: Wei Junfeng; Tang Weidong; Wen Chunming; Huang Longdian
Addresses: College of Electronic Information, Guangxi University for Nationalities, Nanning, Guangxi, China ' Guangxi Key Laboratory of Hybrid Computation and IC Design, Guangxi University for Nationalities, Nanning, Guangxi, China ' College of Electronic Information, Guangxi Key Laboratory of Hybrid Computation and IC Design, Guangxi University for Nationalities, Nanning, Guangxi, China ' Guangxi Yinglun Information Technology Corp., Ltd., Nanning, Guangxi, China
Abstract: Peak detection methods are widely used in various areas. This study introduces a peak detection method used for seed cutting. The envelope of signal in discrete time domain was calculated by Hilbert transform. In order to increase the reliability in noise condition, the envelope was enhanced by utilising instantaneous frequency of signal, and then a multi-rule joint search algorithm marks the peaks, which stands for the location of a node. The proposed method worked in worse SNR condition in the simulation, and was verified in the experiment of detecting sugarcane nodes. Local maximums of accelerometer data are marked, showing the position of node rings on the sugarcane surface. The position data will be used for an action signal in cutting machines, or be used for the analysis of crop growth.
Keywords: discrete Hilbert transform; instantaneous frequency; peak detection; sugarcane seed cutting.
International Journal of Computer Applications in Technology, 2021 Vol.67 No.2/3, pp.149 - 158
Received: 19 Oct 2020
Accepted: 14 Jan 2021
Published online: 17 Mar 2022 *