基于声波的玉米含水量测定.pdfVIP

Vo l. 37 No.3 第 37 卷第3 期 吉林大学学报(工学版) 2007 年 5 月 Journal of Jilin University (巴ngineering and Technology Edition) May 2007 基于声波的玉米含水量测定 孙永海1 ，何小平1 ，孙瑜2 (1.吉林大学生物与农业工程学院，长春 130022; 2. 吉林大学电子科学与工辍学院，长春 130012) 摘 襄:利用声音传感器采集玉米籽粒流从高处落到相同状态玉米堆时发出的撞击声音，对声 音信号进行滤波预处理，提取声波信号强度、功率谱能景、谱峰值等特征值来描述信号。应用 多元线性回归、二项式回归罪口神经网络等方法进行全面分析。试验结果表明:声波信号强度功 率谱能量及谱峰值等特征值与玉米籽粒含水景之间存在较强的相关性，其中信号强度相关性 最大。利用这些特征能够准确地测定籽粒含水量，其中二项式回归模型最接近实测值。 关键询:食品检测技术;含水量测定;声学分析;玉米含水量 中阁分类号:TS210. 7 文献标识码:A 文章编号:1671伽5497(2007)03ω0726-05 Corn moisture measurement based on acoustic analysis Sun Yong-hai 1 ,He Xiao-ping 1 ,Sun Yu2 (1. College 01 Biological and Agricult剖ral Engineering ,]ilin University , Changchun 130022 , China , 2. College 01 Electronic Science and Engineering ,]ilin University ,Changchun 130012 , China) Abstract: A method to measure the corn moisture by acoustic analysis was proposed. The sound wave was acquired by a computer system equipped with a sound sensor. After the filtration of noise from the sound signals , the features , such as the acoustic intensity , spectrum energy and peak value were calculated to de日cribe the sound waves. Multiple linear regression. binominal regression and neural network methods were applied in the analyses. Results show that there exist prominent relationships between these features and the moisture content , and among these features acoustic intensity is the most significant one. Using these relationships the corn moisture content can be accurately measured. 丁he predicted data by the binominal regression model is the closest one to the real data. Key words: food i