Flaw patterns characterization in material images using linear coorelation method

Kumari, Aayushi and Kedia, Priyanshu and Mandal, Eisha and Mandal, S K (2017) Flaw patterns characterization in material images using linear coorelation method. Journal of Metallurgy and Materials Science, 59(2) . pp. 67-74.

Restricted to NML users only. Others may use ->



The structural pattern of any material can be analyzed and predicted using image processing technique. Image processing toolbox available in MATLAB provides us a noble way to perform manipulation on an images. This is done to enhance their features for better representation useful for pattern analysis. Use of image processing technique in the area of structural analysis of material image is relatively unexplored. It has been observed that the properties and attributes of material varies with many factors like temperature, pressure, etc. The variation due to temperature dominates more often. This variation in properties with temperature can be exploited to predict properties at certain different temperature.For proper designing and quality control of the material, it is vital to perform the analysis of microstructure of the material. Image processing proves to be the best tool for performing this analysis which provide various algorithm to aid the prediction process giving maximum accuracy. In this paper we aim to extract existing features from the material image, study their variation with temperature and derive new features to predict the properties at unknown temperature. Any image in MATLAB can be depicted in the form of a matrix where different numbers symbolize variation in color intensity. Also an image can be of various types be it binary image, Grayscale image, RGB image etc. For Binary image the values of pixel can be 0 for black and 1 for white, for a grayscale image 0 represents black color while 255 represents white color and the pixel value can take any value from 0 to 255. For RGB image three dimensional matrix, each dimension representing one of the three colors with 8 bit pixel value. The image with different pixel values representing digitized values representing different colors is call a digital image. When we process a digital image using computers and algorithms to make use out of it then this is called digital image processing. Pattern recognition, at particular physical condition, is one of the application of Digital Image Processing and its use can be extended to predicting the value of the pixels at different physical condition.

Item Type:Article
Official URL/DOI:http://eprints.nmlindia.org/7749
Uncontrolled Keywords:Image processing, Pattern recognition, Structured analysis.
Divisions:Material Science and Technology
ID Code:7749
Deposited By:Sahu A K
Deposited On:12 Feb 2018 12:27
Last Modified:12 Feb 2018 12:31
Related URLs:

Repository Staff Only: item control page