High Resolution three-dimensional (3D) reconstruction of several proteins has been achieved from two-dimensional (2D) crystals by electron crystallography structure determination. However, for badly ordered 2D crystals, especially non-flat crystals, Fourierfiltering based methods fail, while single particle processing approaches can produce reconstructions of superior resolution by aligning particles in 3D space. We have investigated a single particle processing approach combined with the crystallographic method to generate images centered on the unit cells of 2D crystal images. The implemented software uses the predictive lattice node tracking in 2dx/ MRC software to extract particles from the microscope images. These particles are then subjected to a local contrast transfer function (CTF) correction. The tilt geometry obtained in the 2dx software is used to initialize the Euler angles, which along with translations are then refined by a single particle processing approach. Finally, iterative transform algorithms, namely the error reduction algorithm and the hybrid input-output algorithm, are applied to retrieve missing information in the previously obtained 3D reconstruction. Compared with conventional single particle processing for randomly oriented particles, the required computational costs are greatly reduced as the 2D crystals restrict the parameter search space. Preliminary results from a 3D reconstruction of the membrane protein GlpF suggest that the iterative transform process improves 3D resolution.