An Adaptive Statistical Approach for Non-Destructive Underline Crack Detection of Ceramic Tiles Using Millimeter Wave Imaging Radar for Industrial Application

• Published in: IEEE sensors journal Vol. 15; no. 12; pp. 7036 - 7044
• Main Authors: Agarwal, Smriti, Singh, Dharmendra
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-12-2015; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Active imaging; Ceramics; Ceramics industry; convolution; crack detection; Cracks; False alarms; Flaw detection; Imaging; Imaging radar; Inspection; millimeter wave; Monitoring; optimization; radar; Radar imaging; Tiles; convolution; millimeter wave; radar; Active imaging; optimization; crack detection
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2015.2469157

Non-invasive automatic crack detection and characterization of packaged ceramic tiles is an integral part of industrial processes. Inefficiencies in quality monitoring of packaged tiles is quite unbearable in terms of manufacturing, time, money, and consumer satisfaction. As a result, there is an increasing demand for the use of smart visual inspection systems to ensure the high quality of products in the manufacturing line. Although different techniques have been proposed by various researchers, accurate monitoring of the packaged ceramic tiles remains particularly challenging in order to cut down number of false alarms, which could, otherwise, result in fiscal or reputation loss to the company. Therefore, this paper explores the possibility of application of millimeter wave-based non-invasive crack detection of packaged ceramic tiles using an ingeniously designed V band (60 GHz) imaging radar system. A statistics-based adaptive algorithm has been proposed for crack detection of packaged ceramic tiles with the main consideration of achieving minimal false alarm. A total of 40 different ceramic tiles with different crack and non-crack configurations were considered. The proposed algorithm was tested and satisfactory performance has been achieved.