摘要
为了满足采收前后对西瓜成熟度无损检测的需要,实现了在田间环境下通过音频功率谱峰值频率检测西瓜成熟度的方法。在信号预处理部分,提出了使用短时能量和过零率提取出打击信号的算法,能够准确判断打击信号的起止点,完整地提取出每次敲击西瓜的音频信号用于下一步处理,并滤除了录制音频中的微小干扰噪音。在检测成熟度部分,通过实验划分了不同成熟度下功率谱峰值频率区间范围,并用该划分的区间检测西瓜的成熟度。经验证,该方法达到了较高的检测准确度。
This paper works on non - destructive acoustic detection method for maturity of watermelon by the frequency of acoustic power spectrum peak value in field environment, which is used for before and post harvest of watermelon. In pre - processing part, Short - Term Energy and Zero - Crossing Rate algorithm is designed to detect the beginning and end of patting signals for removing the small noises that affects the detection results. This method cuts out every single patting signal for processing in the next step and filters out small noises in the recording signals. In the maturity detection part, the frequency intervals of power spectral peak, which relates to different maturities, are decided by experiment and used to detect the maturity of watermelons. The method in this paper is tested to achieve high detection accuracy.
出处
《农机化研究》
北大核心
2009年第8期150-152,155,共4页
Journal of Agricultural Mechanization Research
基金
河北农业大学非生命学科和新兴学科科研发展基金项目(2007)
关键词
西瓜
成熟度
短时能量
过零率
功率谱峰值频率
watermelon
maturity
short - term energy
zero - crossing rate
the frequency of power spectral peak