Image-Based Process Monitoring Using Low-Rank Tensor Decomposition
Image-Based Process Monitoring Using Low-Rank Tensor Decomposition
复制标题
DOI：
10.1109/tase.2014.2327029
复制
发表时间：
2015-01-01
影响因子：
5.6
通讯作者：
Shi, Jianjun
中科院分区：
文献类型：
--
作者：
Yan, Hao;Paynabar, Kamran;Shi, Jianjun
由于图像和视频传感器可以捕获丰富的过程信息，因此它们越来越多地部署在复杂的系统中。因此，图像数据在制造过程、食品工业、医疗决策和结构健康监测等不同应用领域的过程监测和控制中发挥着重要作用。由于彩色图像具有复杂的数据特性，包括高维性和相关性结构（即时间、空间和光谱相关性），现有的过程监控技术无法充分利用彩色图像的信息。本文提出了一种新的基于图像的过程监控方法，能够处理灰度和彩色图像。所提出的方法用张量对图像数据的高维结构进行建模，并采用低秩张量分解技术来提取使用多元控制图监测的重要监测特征。此外，本文还展示了不同低秩张量分解方法之间的解析关系。通过广泛的模拟和钢管制造过程中的案例研究，对所提出的方法在快速检测过程变化方面的性能进行了评估，并与现有方法进行了比较。从业人员注意事项-本文受钢管制造中燃烧监测问题的启发，重点关注开发基于图像数据的有效过程监测方法。由于彩色图像数据的高维性和复杂的相关结构，现有的过程监控技术无法充分利用彩色图像的信息。本文通过提取基本的监测特征来解决这个问题，同时考虑彩色图像的空间和光谱相关性。这是通过使用各种低秩张量分解方法以及多元控制图来完成的。所提出的方法可以形成计算机辅助在线监控系统，用于自动检测过程中的失控情况。使用仿真，比较了各种场景下所开发方法的性能。这可以为从业者提供有用的指导，以选择合适的基于图像的过程监控方法。在未来的研究中，我们将研究基于图像的故障诊断技术的发展，该技术可以与本文提出的过程监控方法相结合。
Image and video sensors are increasingly being deployed in complex systems due to the rich process information that these sensors can capture. As a result, image data play an important role in process monitoring and control in different application domains such as manufacturing processes, food industries, medical decision-making, and structural health monitoring. Existing process monitoring techniques fail to fully utilize the information of color images due to their complex data characteristics including the high-dimensionality and correlation structure (i.e., temporal, spatial and spectral correlation). This paper proposes a new image-based process monitoring approach that is capable of handling both grayscale and color images. The proposed approach models the high-dimensional structure of the image data with tensors and employs low-rank tensor decomposition techniques to extract important monitoring features monitored using multivariate control charts. In addition, this paper shows the analytical relationships between different low-rank tensor decomposition methods. The performance of the proposed method in quick detection of process changes is evaluated and compared with existing methods through extensive simulations and a case study in a steel tube manufacturing process.Note to Practitioners-This paper, motivated by the problem of combustion monitoring in steel tube manufacturing, focuses on the development of effective methods for process monitoring based on image data. Existing process monitoring techniques cannot fully utilize the information of color images due to the high-dimensionality and complex correlation structure of such data. This paper addresses this problem by extracting essential monitoring features, while considering the spatial and spectral correlation of color images. This is accomplished by using various low-rank tensor decomposition methods along with multivariate control charts. The proposed approach can lead to a computer-aided online monitoring system for automatic detection of out-of-control situations in a process. Using simulation, the performance of the developed methods is compared under various scenarios. This can provide practitioners with useful guidelines for selecting an appropriate method for image-based process monitoring. In future research, we will study the development of image-based fault diagnosis techniques that can be integrated with the process monitoring approaches proposed in this paper.