This thesis addresses the challenge of 3D surface reconstruction using scanning electron microscopy. Scanning electron microscopes find extensive applications in various fields, including engineering, biology, and geology. However, they inherently provide only 2D images. This thesis proposes a method for comprehensive 3D reconstruction, divided into two phases, of the surface of individual microscopic specimens observed through this microscope. In the initial phase, high-quality partial reconstructions are generated from approximately one hundred images obtained by rotating the specimen 360°. The geometric reconstruction approach with precise autocalibration, previously developed in the laboratory (Pollen 3D), is employed.In the subsequent phase, the partial reconstructions are aligned using the ICP algorithm and then integrated to construct a 3D point cloud, which is further converted into a mesh suitable for printing. Along side the 3D reconstruction process, protocols for sample handling within the vacuum chamber are defined, laws controlling centering and autofocus are developed for capturing input images, and methods for filtering 2D images and cleaning point clouds are devised. The developed methodologies were validated using Zeiss and Thermofisher microscopes across various specimens,including pollen grains, turning tools, a foraminifer, and a shark tooth. These reconstructions hold potential utility in educational contexts, preservation efforts for delicate samples, and metrological applications.