在大规模生产太阳能电池板过程中,由于生产工艺的影响,部分电池板表面会产生颜色深浅不同的花纹(又称为晶花)。用户常常要求厂家对不同花纹的电池板进行分类供应。为此,提出了一种人工智能分类识别系统。系统以太阳能电池板的表面花纹深浅程度为分类依据,对太阳能电池板进行分类识别。系统首先使用局部二值模式(LBP)算子作为分类特征,将电池板分为“有晶花”和“无晶花”两类,然后使用局部对比度作为分类特征,对 “有晶花”一类细分为“强晶花”和“弱晶花”两类。为了满足生产线快速、准确分类的要求,系统使用了BP神经网络作为分类器。实验结果表明,分类系统速度快、准确率高,能够满足实际的生产线需求。

In the process of mass production in solar panels, different patterns which are dark or light will be formed on the surface of the panels due to the production technology. An automatic classification system is needed by the factory to improve work efficiency, so an artificial intelligence identification system as a classifier is proposed to meet this demand. The solar panels will be classified by the system using the surface pattern of solar panels as a reference. Firstly, local binary pattern (LBP) operator is used as classification features, the samples of solar panels will be divided into two categories which are separately named With crystal and No crystal. And then, the local contrast is taken as the classification feature, and the with crystal sample are also divided into two categories such as dark crystal and light crystal. In order to meet the rapid and accurate demands of industrial production, back propagation (BP) neural networks are adopted in auto-recognition as a classifier. Experimental results show that the classification system has fast speed and high accuracy rate, which can meet the demands of actual industrial production line.