蛋壳破损自动检测模型研究

为了建立蛋壳破损检测模型，试验采用敲蛋装置与声检测控制器组成的计算机蛋破损检测系统获取被检蛋声音信号数据。通过对蛋声音信号的功率谱分析而得到反映蛋壳破损特征的参数：功率谱面积的平均值x₁，最大功率谱面积与最小功率谱面积的差值x₂，X轴方向上质心的平均值x₃，X轴方向上质心最大值与其最小值的差x₄，Y轴方向上质心的平均值x₅，Y轴方向上质心最大值与其最小值的差x₆，共振峰频率的最大值x₇和共振峰频率的最大值与其最小值间的差值x₈；再通过Bayes原理建立与蛋壳破损特征参数相关的蛋壳破损模型。检验结果表明模型准确率达到92%。

Model for automatic detection of eggshell crack

In order to establish models for cracked eggs and intact eggs, a detecting system that was composed of an automatic impact device and a response sensor controller was used. Characteristic parameters such as the average area of power spectrum(x₁), the difference between the maximum and minimum values of area of the power spectrum (x₂), the average of the x-coordinate of centriod (x₃), difference between the maximum and minimum values in the x-coordinate of centroid (x₄), the average of the y-coordinate of centriod (x₅), the difference between maximum and minimum values in the x-coordinate of centroid (x₆), the average peak resonant frequency (x₁) and difference between the maximum and minimum values in peak resonance frequency (x₈), were achieved for differentiating cracked eggs and intact eggs according to the analysis of their power spectra in this research. Bayes theorem was applied to develop discrimination models for cracked eggs and intact eggs. The test showed that the average accuracy of detection models was about 92%.