论文部分内容阅读
介绍了一个苹果内部品质自动检测系统的原理性样机 ,该样机在计算机控制下基于介电特性实现苹果品质检测。待检测苹果的介电参数通过两只电容传感器和 L CR(电感 -电容 -电阻 )测量仪进行采样并输入样机内的控制计算机。选择等效电容 (介电参数之一 )作为苹果等级分类变量。对于一批待检测的苹果 ,其最佳分类阈值可通过对抽选出的合格和等外样本的在线学习获得。试验结果表明 ,该样机可获得良好的分类效果。在待检苹果直径为 7.0 cm~ 8.5 cm ,坏损部分的直径大于 2 cm,电容传感器工作频率为 1k Hz时 ,以学习得到的最佳分类阈值进行分类 ,分级正确率大于 80 % ,分级效率达到 2 .5 s分选一个苹果 ,接近人工分选效率
The principle prototype of an apple internal quality automatic detection system is introduced, which realizes apple quality inspection based on the dielectric properties under the control of a computer. The dielectric parameters of the apple to be tested are sampled by two capacitive sensors and a LCR (Inductance-Capacitance-Resistance) meter and input into the control computer in the prototype. Select the equivalent capacitance (one of the dielectric parameters) as the apple rating classification variable. For a group of apples to be tested, the best classification threshold can be obtained by online learning of the selected qualified and out-of-sample samples. The experimental results show that the prototype can get good classification results. When the diameter of the apple to be inspected is 7.0 cm ~ 8.5 cm, the diameter of the damaged part is greater than 2 cm, and the operating frequency of the capacitive sensor is 1 k Hz, the classification threshold is classified according to the best classification threshold, the classification accuracy is more than 80% Achieve 2.5 s sorting an apple, close to manual sorting efficiency