生物视觉模型的太阳能电池板裂纹检测  

Biologic Visual Model for Inspection of Solar Cell Crack

作　　者：方帅[1] 李斌[1] 何祥浩 

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009 [2]安徽方圆机电股份有限公司,安徽蚌埠233010

出　　处：《小型微型计算机系统》2012年第8期1868-1872,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60705015)资助

摘　　要：太阳能作为一种具有广阔前景的新能源,在太阳能电池制造产业对太阳能电池板裂纹检测具有重要的生产实践意义.过去的硬件方法效率低且容易使太阳能电池受损,而软件方法主要是通过经典的图像处理方法进行检测,但对复杂背景的裂纹效果一般.通过研究模拟腹侧视觉通路的工作原理的T.Serre生物视觉标准模型,使用Tomoyuki的Percolation-Based图像处理方法指导特征模板选取,提出生物视觉模型的改进方法.并通过对一系列采集的太阳能电池板图像的检测实验和对比实验,证明了该方法的有效性.Solar energy as a kind of promising new energy, inspection of solar cell crack has important practical and productive significance in solar cell manufacturing industry. The hardware methods are inefficient in the past, and easily damaged solar cells. But software methods in the past mainly through the classical image processing methods for detection, and have the general effect in complex background. We propose an improved method of biological vision model, by simulating the working principle of the ventral visual pathway of T. Serre standard model of biological vision, and using Tomoyuki the Percolation-Based Image Processing to guide the selection of feature template. Through the detection experiments and comparative test on a series of collected solar cell images, demonstrated the effectiveness of our method.

关 键 词：太阳能电池板 图像处理 裂纹检测 标准模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]