Prediction of grain boundary of a composite microstructure using digital image processing: A comparative study Sweta Rani Biswal a,⇑ , Tapasmini Sahoo b , Seshadev Sahoo a a Department of Mechanical Engineering, ITER, Siksha ‘O’ Anusandhan (Deemed to be University), Bhubaneswar, India b Department of Electronics and Communication Engineering, ITER, Siksha ‘O’ Anusandhan (Deemed to be University), Bhubaneswar, India article info Article history: Received 29 July 2020 Received in revised form 13 September 2020 Accepted 16 September 2020 Available online xxxx Keywords: Hybrid composite Microstructure Grain size Edge detection operator Digital image processing abstract In recent years, due to its balance mechanical and tribological properties, hybrid materials have occupied a dominant position in the automotive industry and aerospace industry. However, in the study of mechanical and tribological properties, it is necessary to perform metallographic characterization of hybrid composites. The discovery of digital image processing technique as a next generation technology for metallographic characterization results in time saving and error free solution. This automation tech- nology includes multiple stages, such as image enhancement. Perform spatial filtering to remove noise, edge detection, and watershed conversion. One of the key factors of digital image processing technique is the application of statistical analysis in edge detection stage. This signify, edge detection stage plays the crucial role in digital image processing technique which provide a good quality microstructure from an unfavorable image. As per our comparative study between the edge operators like Sobel operator, Robert operator, Prewitt operator and Canny operator, we recommend the results of canny edge detectors for watershed transformation to calculate the grain edge to find the average grain size and results of sta- tistical analysis method validate the same. The results also confirmed the formation of distinct edge of grain after watershed transformation. Ó 2020 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Recent Advances in Mechanical Engineering Research and Development. 1. Introduction Over the past few years, many researchers found automated computerized image analysis systems that can execute the microstructural images of different materials in a simpler and much effective way. This provides a modern technical approach to process the optical microscopic structure, scanning electron microscopic structure, and transmission electron microscopic structure. Automated computerized image analysis systems pro- cess the image and find the grain size using various statistical methods. This also helps in the detection of different phases and dislocation [1–4]. Some researcher has explained that Computer- based image processing technology has a greater impact than the traditional manual process which are generally time consuming and erroneous. These technologies not only make the processing of microstructure simpler and time-saving but also make it predic- tive and accurate by reducing errors. Advancement in Computer- based image processing technology processed many testing sam- ples at once. The improvement in the detection level gives more effective results [5,6]. This technology not only changes the vague digital microstructure but also processed them through filters, edge detection operators, and watershed transformation into a clear and acceptable digital form. These processing steps of the dig- ital image processing technique play a vital role in the identifica- tion of the grain size for many materials. Researchers also found that digital image processing is advantageous, but to facilitate a proper digital image processing technique is still under research. In this paper, we intended to find a proper solution for the selection of a digital image processing technique for the material scientist. Here, we have emphasized the role of edge detectors. We have taken all the available edge detectors like Sobel Operator, Robert Operator, Prewitt Operator, and Canny Operator for the pro- cessing of the digital image of Al-Al 2 O 3 -WS 2 based hybrid compos- ite at different composition. Aluminum-based hybrids composites are those composites who consist of multiple reinforcements https://doi.org/10.1016/j.matpr.2020.09.559 2214-7853/Ó 2020 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Recent Advances in Mechanical Engineering Research and Development. ⇑ Corresponding author. E-mail address: swetaranibiswal@soa.ac.in (Sweta Rani Biswal). Materials Today: Proceedings xxx (xxxx) xxx Contents lists available at ScienceDirect Materials Today: Proceedings journal homepage: www.elsevier.com/locate/matpr Please cite this article as: Sweta Rani Biswal, T. Sahoo and S. Sahoo, Prediction of grain boundary of a composite microstructure using digital image pro- cessing: A comparative study, Materials Today: Proceedings, https://doi.org/10.1016/j.matpr.2020.09.559