摘要:
近年来图像处理技术在工业产品的自动化检测中得到了广泛的重视和应用，针对电池板尺寸测量和边缘缺陷检测，本文首先提出通过局部边缘像素拟合电池板的几何形状来计算电池板的表面尺寸，然后使用逆向遍历法检测指定边的像素集合，最后通过边缘像素的曲率计算和阈值处理实现边缘缺陷的检测。实验表明，本文中的算法能够准确并快速地实现电池板的尺寸测量，并且通过曲率定位边缘缺陷的新方法效果显著。
Abstract: Image processing techniques have attracted much attention in the field of automatically detection of industrial products. For the size measurement and defect detection of solar cell edges, fitting geometrical shape of the solar cell by local edge pixels is proposed to measure the sizes. Then a reverse traverse technique is used to detect the edge pixels. Finally, a curvature computation and a threshold processing are proceeded to detect the edge defects. Experimental results indicate that the proposed algorithm can measure the sizes accurately and quickly, and the new method to detect the defects of solar cell edge by curvature works effectively.