%0 Journal Article
%T 基于RPCA的铝管表面缺陷检测<br>Surface Defect Detection of Aluminum Tube Based on RPCA
%A 樊浩然
%J Artificial Intelligence and Robotics Research
%P 237-245
%@ 2326-3423
%D 2020
%I Hans Publishing
%R 10.12677/AIRR.2020.94027
%X 工业上铝管在装配使用之前要检测铝管是否存在缺陷，本文采用一种基于RPCA的铝管表面缺陷检测，主要检测铝管表面的擦伤和压坑缺陷。为保证铝管图像之间线性相关，要先对图像预处理，得到一组能够用RPCA (Robust Principal Component Analysis)算法处理的铝管图像。本文采用RPCA算法对一组分辨率为250 × 75的8张铝管图像进行低秩稀疏分解，运用非精确拉格朗日乘子法求解RPCA模型，得到的稀疏图像即为缺陷图像，对缺陷图像进行识别。通过实验证明上述方法在铝管表面缺陷检测上具有可行性。<br />
In industry, before the aluminum tube is assembled and used, it is necessary to detect whether there are defects in the aluminum tube. In this paper, an RPCA based surface defect detection method is used to detect the scratch and indentation defects on the surface of the aluminum tube. In order to ensure the linear correlation between the aluminum tube images, it is necessary to pre-process the images to get a group of aluminum tube images which can be processed by RPCA algo-rithm. In this paper, the RPCA algorithm is used to decompose a group of eight aluminum tube im-ages with a resolution of 250 × 75. The RPCA model is solved by using the Imprecise Augmented Lagrange Method. The sparse images are defect images, and the defect images are identified. The experimental results show that the above method is feasible in the surface defect detection of alu-minum tube.
%K RPCA，缺陷检测，预处理，非精确拉格朗日乘子法，图像识别<br>RPCA
%K Defect Detection
%K Pretreatment
%K Imprecise Augmented Lagrange Method
%K Image Identifica-tion
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=38509