Automatic Grain Size Detection and Classification of Metals

Author Name(s): Vinayak Bairagi, Mayuri Rathod, Tushar Shirsath
Author Email: vbairagi@yahoo.co.in

Abstract

Metallography is the study of the microstructure of metallic alloys. The integrity of metals in vital in the case of any manufacturing industry as the durability is dependent on quality of metal. Microscopic analysis is essential part of modern manufacturing process. Micro-structural Inclusions measurement is routinely required process for maintaining quality. The most important physical properties of particulate samples are particle size and grain size. The diameter of individual grains of sediments is called as grain size. Gran size determination is actually depends upon the manual process which leading to time-consuming and errors are occur at the time of determination. The main objective to propose this system is to automatically analyze the material according to their size/shape and grain sizes. Fuzzy logic model is a classifier used to detect edges and corners of grains available in the input image. The system categories the metal using its grain size. The obtained results are compared with ASTM standards and other reported results in survey. The proposed system is useful in metal industry which are depends upon size and shape of the grains.

Introduction

Metallography is important part of study related to the microstructure of metallic alloys. It can be stated that the scientific discipline of observing and analyzing the spatial distribution of the sediments, inclusions or phases present in metallic alloys. By extension, these similar principles can be applied to the further characterization of any material or alloy. Different techniques are used to reveal the micro structural characteristics of alloys [1]. An ultramicroscopic is a microscope with illumination system that allows viewing of tiny particles even up to size of micro meter level. Microscopic analysis is essential part of modern manufacturing process in metal industry. The quality, durability, and life cycle of a final product depend on the attribute and characteristics of the material with fitting tolerances. A defect in single sediment can lead to breakdown the entire end product of system. Micro-structural Inclusions measurement is routinely required process for maintaining quality. The most important physical properties of particulate samples are particle size and grain size. The main objective to propose this system is to automatically analyze the material according to their size/shape and grain sizes. Grain size is diameter of individual grains of sediments. Grains are having internal boundaries to define grain size of metal. Grain size analysis acts an important part in the research of metallic material. From grains analysis we can get the information related to material’s properties like yield strength, tensile strength and elongation which have effective impact on material’s properties [22]. Grain size is “Diameter of individual particles”. The methods of grain analysis are relying on the manual processes but it is time consuming and errors are liable to occur. So with the development of image processing technologies the DIP (Digital Image processing) and pattern recognition technologies can be used as tools for calculating the grain size automatically. Image-pro plus, image tool are widely used in metallographic process. This technique reduces manual workloads and improves the efficiency of analysis but it is difficult to identity grain boundaries in the image frame with low contrast and bad boundaries detected [22-27]. So we overcome these problems with this system using fuzzy logic.

Conclusion

In this paper, we have presented a new combination of algorithms to do material analysis. To ensure the consistent changes, a material analysis is proposed approach to classify varieties of metal alloys of different types as shown in Table No.2&3. We have compared the proposed results with existing work and also with ASTM standard. We have introduces material analysis/Grain size analysis according to their size/shape by using two new different methods like direct formulae based method and intercept method based on fuzzy logic. The output of proposed system, the grain size of particular metal is detected automatically. Experimental results on material analysis/Grain size analysis have shown the consistency of the proposed methods and its application. The results can also be improved by use of 3D images at input image. The use of pre-processing will be useful in detecting edges correctly.

820 total views, no views today

Download PDF File

About the author: admin