Real-Time Detection of Infrared Profile Patterns and Features Extraction

The pressing demand to improve quality of manufactured products requires the use of the latest technologies in order to enhance the control systems that adjust manufacturing parameters. Computer vision inspection and control are already standard technologies which are frequently used to improve the quality of manufactured products. Recently, due to the availability of fast and affordable infrared acquisition devices, computer vision beyond the spectrum is also becoming an essential technology for quality control and improvement. For example, during steel strips manufacturing, uneven temperature across the width of the strips during rolling generates defects, due to differences in the contraction of the longitudinal fibers that make up a strip. Detecting the infrared profile pattern of the strip which is being manufactured makes it possible to use this information to modify the manufacturing parameters to compensate for the temperature differences in the strip (Gonzalez et al., 2002). This work proposes a robust method to detect infrared profile patterns in real-time. The proposed method is based on the acquisition and processing of infrared profiles using an infrared line scanner. The detection of infrar ed patterns, and the change of pattern which occur during manufacturing, need to be carried out online with the production process in order to use this information to enhance the control systems during manufacturing. The method proposed to detect these patterns in real-time is based on the segmentation of the stream of infrared profiles acquired from the infrared line scanner. The segmentation aims to find regions of homogeneous temperature, that is, regions formed by a set of adjacent profiles which have a similar temperature pattern. The proposed method to segment infrared images into regions of common temperature patterns is by means of boundary detection, which, in this case, is accomplished through edge detection. The first step of the segmentation is the calculation of the gradient, which is obtained as the result of the convolution of the image with a gradient operator. Two different gradient operators, Gaussian and difference, are evaluated to test which one is best suited to solve the current problem. The next step is the projection of the gradient, which simplifies the thresholding that must be carried out to eliminate noise from the gradient. Once the projection of the gradient is available, it is thresholded. The objective of the thresholding is to differentiate noise from real edges. An edge is found when there is data in the projection over the

[1]  Paul L. Rosin,et al.  Evaluation of global image thresholding for change detection , 2003, Pattern Recognit. Lett..

[2]  Peter H. A. Sneath,et al.  Numerical Taxonomy: The Principles and Practice of Numerical Classification , 1973 .

[3]  Peng-Yeng Yin,et al.  Pattern Recognition Techniques, Technology and Applications , 2008 .

[4]  Milton Abramowitz,et al.  Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables , 1964 .

[5]  Rita Cucchiara,et al.  Tuning Range Image Segmentation by Genetic Algorithm , 2003, EURASIP J. Adv. Signal Process..

[6]  Hui Zhang,et al.  Image segmentation using evolutionary computation , 1999, IEEE Trans. Evol. Comput..

[7]  D. E. Goldberg,et al.  Genetic Algorithms in Search , 1989 .

[8]  Alden H. Wright,et al.  Genetic Algorithms for Real Parameter Optimization , 1990, FOGA.

[9]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[10]  Chang Wook Ahn,et al.  On the practical genetic algorithms , 2005, GECCO '05.

[11]  Goldberg,et al.  Genetic algorithms , 1993, Robust Control Systems with Genetic Algorithms.

[12]  R. Usamentiaga,et al.  Application of thermographic analysis for the detection of longitudinal defects in cold mills , 2001 .

[13]  Bir Bhanu,et al.  Adaptive image segmentation using a genetic algorithm , 1989, IEEE Transactions on Systems, Man, and Cybernetics.

[14]  Hyun Seung Yang,et al.  Robust image segmentation using genetic algorithm with a fuzzy measure , 1996, Pattern Recognit..

[15]  R. Mukundan,et al.  Some computational aspects of discrete orthonormal moments , 2004, IEEE Transactions on Image Processing.

[16]  David G. Stork,et al.  Pattern Classification , 1973 .

[17]  Bülent Sankur,et al.  Survey over image thresholding techniques and quantitative performance evaluation , 2004, J. Electronic Imaging.

[18]  R. Sokal,et al.  Numerical Taxonomy: The Principles and Practice of Numerical Classification. , 1975 .

[19]  Philippe Andrey,et al.  Selectionist relaxation: genetic algorithms applied to image segmentation , 1999, Image Vis. Comput..

[20]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[21]  Y. J. Zhang,et al.  A survey on evaluation methods for image segmentation , 1996, Pattern Recognit..

[22]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[23]  Nicholas J. Radcliffe,et al.  Forma Analysis and Random Respectful Recombination , 1991, ICGA.

[24]  Shigeo Abe DrEng.  Pattern Classification , 2001, Springer London.