The detection of defects in solar cells based on machine vision has become the main direction of current development, but the graphical feature extraction of micro-cracks, especially cracks with complex shapes, still faces formidable challenges due to the difficulties associated with the complex background, non-uniform texture, and poor contrast between crack defects and background. In this paper, a novel detection scheme based on machine vision to detect multi-crossing cracks for multi-crystalline solar cells was proposed. First, faced with periodic noise, we improved the filter method in the frequency domain and eliminated the background interference of fingers by filtering out the periodic noise while retaining the integrity of the crack signal. To address the anisotropy of multi-crossing cracks, we designed a special grid-shaped, convolution kernel filter to accurately extract crack features at low contrast and in the presence of a complex textured background. Finally, to address the missing features from the central region of multi-crossing cracks, we designed a method based on the orientation information of mask pattern to implement feature reconstruction for the central region of the crack. The experimental results showed that, compared to other crack detection methods, the strategy designed herein exhibited a better detection performance and stronger robustness.

基于机器视觉的太阳能电池缺陷检测已成为当前发展的主要方向，但是由于复杂背景的相关困难，微裂纹（尤其是形状复杂的裂纹）的图形特征提取仍然面临着严峻的挑战，质地不均匀，裂纹缺陷和背景之间的对比度差。提出了一种基于机器视觉的多晶太阳能电池多交叉裂纹检测方案。首先，面对周期性噪声，我们在频域上改进了滤波方法，并通过滤除周期性噪声，同时保留了裂纹信号的完整性，消除了手指的背景干扰。为了解决多次交叉裂纹的各向异性，我们设计了一种特殊的网格形 卷积核滤波器可以在低对比度和复杂纹理背景下准确提取裂纹特征。最后，为了解决多交叉裂纹中心区域的缺失特征，我们设计了一种基于掩模图案方向信息的方法，对裂纹的中心区域进行特征重建。实验结果表明，与其他裂纹检测方法相比，本文设计的策略具有更好的检测性能和更强的鲁棒性。我们基于掩模图案的方向信息设计了一种方法，以实现裂纹中心区域的特征重建。实验结果表明，与其他裂纹检测方法相比，本文设计的策略具有更好的检测性能和更强的鲁棒性。我们基于掩模图案的方向信息设计了一种方法，以实现裂纹中心区域的特征重建。实验结果表明，与其他裂纹检测方法相比，本文设计的策略具有更好的检测性能和更强的鲁棒性。