基于多尺度小波变换的鸡蛋散黄检测方法研究

Study on Detection of Scattered Yolk Eggs Based on Multi-Scale Wavelet Transform Method

摘要:

为建立一种快速有效的无损检测鸡蛋散黄的方法，构建了基于磁致伸缩振子扫频式振动的鸡蛋散黄检测系统，通过对采集的鸡蛋振动音频信号进行多尺度小波变换分析，找到新鲜蛋与散黄蛋的音频信号差异性，基于此提取合适的特征值，并分别构建基于BP神经网络、RBF神经网络和Hopfield神经网络的鸡蛋散黄检测模型加以比较。实验中，对300枚鸡蛋进行检测（训练集200枚，测试集100枚），结果表明，Hopfield神经网络对新鲜蛋和散黄蛋的识别效果最好，测试集中新鲜蛋和散黄蛋的识别率均达到98%，且检测每枚鸡蛋的时间为31.6 ms。研究表明，利用磁致伸缩振子扫频振动未知品质鸡蛋，再通过小波变换分析，并结合Hopfield鸡蛋散黄检测模型检测鸡蛋散黄是可行的。

Abstract:

Ferric chloride is widely used in the wastewater treatment. The mechanism study on the flocculating of ferric chloride and the recent progress were discussed in this paper. Four typical mechanisms were illustrated including charge neutralization，adsorption bridging，sweep flocculation and compressing thickness of electric double layer. The current research progress was reviewed and intensive study was suggested on floc structure，conformation，zeta potential and the synergistic effect of ferric chloride.