基于机器视觉的太阳能电池颜色自动分选研究

Study of Color Sorting Automatically in Solar Cells Based on Machine Vision

针对硅晶太阳能电池颜色差异,首先设计了一套颜色分选方案,构建了图像采集与自动化单元;并提出了基于彩色机器视觉原理的颜色分选方法,通过对图像分块直方图的计算与比较建立颜色样本库与待检测目标之间的相似关系,并以此作为颜色分选的依据。最后将以往计算整体灰度直方图差异来进行颜色分选的方法相比较,证明了所提方法能够克服色差对分选结果的影响。实验证明,所提方法在配合系统工作中具有足够的快速性、准确性和稳定性,能够满足实际生产需求。

For the quality problem caused by color difference in crystalline solar cells,a project of sorting color is designed. Firstly,image acquisition and automation system is constructed. Then,a method of sorting color based on colorful machine vision is proposed: the similarity betw een the normal cell and the detected one is calculated by segmenting image and the block gray histogram and is used to check the color kind.Lastly,It has been proved that compared w ith calculating the gray histogram entirely,the proposed performs better w hen it comes to cells w ith color difference. Experiments show that the proposed has sufficient rapidity,accuracy and stability,it is possible to meet the actual production needs.