A Quantifiable Detection of voids/Pores and Cracks in Scanning Electron Microscopic images using Machine Learning Techniques

Recently, Image Processing and analysis is an important in all fields such as science, mechanical and medical field. SEM (Scanning Electron Microscopic) read images and extracts the useful information from the samples. The mechanical expert’s analysis the samples in qualitative. They are missing many pores/voids and cracks in the SEM images so the strength of material is not identified properly. By taking this issue, a quantitative analysis is done by so can identify exact pores and cracks present in the SEM image. Here applied machine learning algorithm such as classification and regression algorithm. It classifies the images in the form of RGB where masking techniques is applied. It can analyse the edges of cracks, pores/voids in SEM image. Linear regression predicts the material strength as strong and weak. Here we used improvised classification and regression algorithm. It helps to predict material strength at initial stage. Hence this algorithm is works for both metal and non-metal Images.