Percutaneous nephrolithotomy (PCNL) is the most commonly used procedure to remove large stones from the human kidney. To improve the speed and safety of PCNL, researchers have been developing an intraoperative navigation system which employs a marker-based registration technique to superimpose CT images on the video stream of a tablet computer. In this paper, we present our work on the fusion of intraoperative X-ray and video aimed at improving the existing system. For the fusion, we used automatic marker-detection algorithms and then processed the data by using thin-plate spline transformation and landmark warping. The evaluation of the technique was performed by testing it on a silicone phantom. The results of the evaluation are very promising, showing a mean geometrical error of 0.9 mm. The aggregated runtime of our algorithms is 0.086 ms, which gives the fusion real-time capability. The fusion enables the detection of organ deformations by the surgeon. Possible future work could include the development of a registration framework to automatically detect organ deformations. Furthermore, algorithms for needle detection in fluoroscopic images would greatly improve intraoperative navigation. The location of the needle could then be superimposed on the tablet screen.