基于压缩感知的微组装焊点质量三维检测方法研究

项目名称： 基于压缩感知的微组装焊点质量三维检测方法研究

项目编号： No.61503128

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 赵辉煌

作者单位： 衡阳师范学院

项目金额： 20万元

中文摘要： 焊点质量检测是保障微组装电子产品质量和可靠性的关键技术。随着电子元件的微型化发展，传统的检测方式由于误差大，检测故障类型有限等原因，难以满足焊点质量检测要求，焊点质量高精度检测方法是亟待解决的重要课题。先对微组装电路模块的三维表面高度点云进行特征分析，研究一种三维点云分割算法，从中提取出焊点的表面三维点云信息；然后，利用焊点表面三维点云的稀疏或局部稀疏特性，根据压缩感知理论，对三维点云数据压缩与重构展开理论推导，建立基于压缩感知的三维点云稀疏模型，并以最小化范数方法为思路，研究三维稀疏模型的快速重构算法；最后，研究一种基于稀疏模型的焊点表面三维点云分类识别方法，实现对焊点质量检测。本项目研究的微组装焊点质量检测新方法，不仅可解决传统方法中存在的缺陷，提高检测精度，还可覆盖更多类型的焊点故障检测。研究结果对保障微组装电子产品质量和可靠性具有重要意义,在微型物体识别研究方面具有重要的理论价值。

中文关键词： 视觉检测；焊点；压缩感知；微电子组装

英文摘要： Solder joint quality inspection is the key technology for protecting the microelectronic assembly products quality and reliability. With the development of microelectronics assembly components, the traditional inspection way which has low accuracy and a limited detection range cannot meet the requirements of solder joint quality inspection. It is urgent to develop a high precision method for solder joint quality inspection. At first, by analyzing the features of 3D point cloud, a 3D point cloud segmentation algorithm is proposed, and solder joint surface 3D point cloud is got. Then, According to the theory of compressed sensing, 3D data sparse representation and reconstruction are researched. And a 3D sparse model based on compressed sensing is established, then its fast solving algorithm based on minimum norm method is researched; At last, with the use of sparse and local sparse characteristic, a recognition method of for 3D point is developed, and the solder joint quality inspection is achieved. A novel method for solder joint quality inspection based on compressed sensing is proposed. Compare with the traditional inspection way, the novel method can not only solve the defects existing in traditional method, and improve the detection accuracy, also cover more types of solder joint failure inspection. The research results have important significance to protect the microelectronic assembly products quality and reliability, and it has an important theoretical value for the micro-object recognition.

英文关键词： Vision inspection;Solder joint ;compressed sensing;Microelectronic assembly