Access Type

Open Access Thesis

Date of Award


Degree Type


Degree Name



Mechanical Engineering

First Advisor

Xin Wu


A novel image processing platform is developed in this thesis for pre-processing the images of two-phase microstructures containing isolated particles based on Matlab®. The two phases in this work represent martensite particles and ferrite matrix in Dual Phase steels that are developed for lightweight automotive body structures. An automated image processing tool is useful for obtaining statistical microstructure characteristics, important for better understanding material's deformation and fracture behavior. However, a generalized commercial software can not automatically realize the microstructural features of special interest, and a user-specified pre-processing tool is needed.

This platform utilizes the Matlab® GUI technique, Matlab image processing functions, fuzzy technology to distinguish martensites from ferrites, optimize the intermediate martensite results, analyze and generate the martensite distribution information and volume fraction. The multiple region thresholding technique is proposed to enhance the image segmentation. A fuzzy linguistic system is presented to utilize experts' experience to remove noise and fill the unexpected holes on the image. The examples show that with this platform the pre-processing on images of two phase microstructures containing isolated particles is more convenient, fast and accurate.