基于改进Canny算法的航天电子装联焊点缺陷检测方法

Research on the Application of Machine Vision Technology in Aerospace Electronic Assembly

摘要:

航天电子装联中焊点的质量对设备的可靠性至关重要,焊接缺陷的检测是确保系统稳定运行的关键。为提升航天电子装联过程中焊接缺陷检测的效率和准确性,研究提出了一种基于改进Canny算法的图像处理方法。研究采用双边滤波技术在平滑噪声的同时保留边缘信息,结合Otsu阈值分割算法来自动确定最佳阈值。通过设置双阈值来确定强边缘和弱边缘,引入Hough变换填补断裂的边缘。实验结果表明,基于改进Canny的系统在各类焊点检测中表现出色,在0.12 mm网板中,正常焊点和桥接焊点的检测的准确率分别为98.89%和98.21%。在0.18 mm网板中,该系统的在少焊和过焊焊点的检测准确率分别为97.56%和98.47%。同时,验证了改进Canny算法在准确率、漏检率和运行时间方面的差异均具有统计学意义。此外,改进Canny系统在计算时间和CPU使用率上优于其他两种算法,表现出该系统在实际应用中的优越性。
关键词:机器视觉；边缘检测；Canny；图像预处理；航天电子装联

Abstract:

The quality of solder joints in aerospace electronic assembly is crucial for the reliability of equipment, and the detection of welding defects is the key to ensuring the stable operation of the system. To improve the efficiency and accuracy of welding defect detection in the aerospace electronic assembly process, an image processing method based on an improved Canny algorithm is proposed. The study adopts bilateral filtering technology to smooth noise while preserving edge information, and combines Otsu threshold segmentation algorithm to automatically determine the optimal threshold. By setting dual thresholds to determine strong and weak edges, Hough transform is introduced to fill the broken edges. The experimental results show that the system based on the improved Canny performs well in various types of solder joint detection. In the 0.12mm mesh board, the accuracy of detecting normal solder joints and bridging solder joints is 98.89% and 98.21%, respectively. In the 0.18mm mesh board, the detection accuracy of the system for low solder and over solder joints is 97.56% and 98.47%, respectively. At the same time, it was verified that the improved Canny algorithm had statistically significant differences in accuracy, missed detection rate, and running time. In addition, the improved Canny system outperforms the other two algorithms in terms of computation time and CPU usage, demonstrating its superiority in practical applications.