基于神经网络的黄瓜等级判别

在瓜果的等级判别中尚存在机器人挑选与人挑选结果不一致的问题。作者基于图像处理技术和神经网络理论开发了一种适用于长型瓜果的判别系统，该系统通过对各对象标准样本的学习即可实现对学习工型瓜果的等级判别，达到一机多用的目的。

NEURAL NETWORK BASED ON CUCUMBER GRADER JUDGEMENT

Judging results are not always the same between robot and human in sorting vegetable and fruits. The reason is that the shape and quality of most agricultural products can not be easily expressed mathematically to give a constant judging standard. In order to upgrade the robot sorting accuracy and minimize the difference between robot and human, a judging system based on image processing and neural network technologies was developed for long shape fruit and vegetable sorting. After training with sample patterns of different objects, the system can accurately sort different kind or variety of long shape fruit and vegetables. As a result, the judging system is suitable for multiple purposes. The experimental results show that the judging accuracy of the developed system increase to 96 7% for cucumber. Moreover, the system is easy to operate and does not require any special skilled operators.