This project performs a 3D reconstruction of a cylindrical asphalt mixture sample from a set images on a Dicom file format. The components of asphalt mixture will be identified through 3-phase segmentation. Additionally, a 3D model of the asphalt mixture reconstruction was developed. The project was implemented on the Python programming language using the open-source libraries Numpy, Scipy, Pydicom, Scikit-learn, Matplotlib and Mayavi. A simulation of the asphalt mixtures aging process is implemented using numerical methods on the mechanical, themical and chemical fields.