基于神经网络分析的鲜蛋破损检测

为了提高禽蛋生产中检测自动化程度,提出一种替代人工检测鲜蛋的自动化方法。该方法根据鲜蛋受到专用仪器轻轻敲击所发出的声音的频谱谱形识别鲜蛋的状态。谱形通过本文改进的BP神经网络进行识别．测试结果表明,本方法比较可靠。

Automatic Detection Damaged Fresh Eggs Based on Neural Networks

An automatic technique of fresh egg detection is developed in this paper to replace the technique by hand. The automatic technique is based on the pattern of the frequency spectrum of the sound emitted by the egg due to a slight impact by an impact device. The pattern of the spectrum, which reflects back the condition of the egg is recognized by BP neural networks improved in this work. Test results indicate that the technique is fairly reliable.