基于短时傅立叶变换的鸡蛋破损检测技术的研究  

Detection of Egg Cracks by Short-Time Fourier Transform

作　　者：平建峰[1] 吴坚[1] 应义斌[1] 

机构地区：[1]浙江大学生物系统工程与食品科学学院,杭州310029

出　　处：《传感技术学报》2009年第7期1055-1060,共6页Chinese Journal of Sensors and Actuators

基　　金：国家科技支撑计划资助(2006BAD11A12);国家自然基金资助(30670535);教育部"新世纪优秀人才支持计划"基金项目(NCET-07-0752)

摘　　要：声学技术是目前应用于鸡蛋破损检测中最具潜力的一门技术,它利用完整蛋和破损蛋在声脉冲响应上的差异性来对鸡蛋进行分级分类。本文所研究的系统中采用前置放大器和A/D转换芯片采集鸡蛋被敲击后发出的声音,对获得的时域信号进行频域分析。相比于传统的傅立叶分析,本文采用了时频分析技术,即利用短时傅立叶变换来分析鸡蛋响应信号,并建立了相应的分级分类模型。实验表明基于短时傅立叶技术的鸡蛋破损检测模型对破损鸡蛋具有很高的识别率,为实际生产过程中鸡蛋的无损检测提供了新的并且更有效的方法。Presently, acoustic technique is the most potent one of techniques applied to detect egg cracks, which classifies eggs through making use of the differences between intact eggs and cracked eggs in impulse response. In this research, preamplifier and A/D conversion chip are utilized to collect signals from the knocked eggs, then the collected time signals are analyzed in frequency field. In comparison to traditional Fourier analysis, this research utilizes time-frequency technique, i.e. short-time Fourier transform, to an- alyze egg response signals, and sets up relevant classification model. Experiments show that this model has high identification rate in detection of cracked eggs, thus this technique provides new and available method for nondestructive detection of eggs in practice.

关 键 词：蛋壳 傅立叶变换 短时 时频分析 

分 类 号：O422.3[理学—声学]