Contrast media is widely used in hospital for a better imaging of CT angiography. However, emergencies (such as needle eruption and capillary hemorrhage) may exist due to the extreme high speed of injection. A video-object-tracking system is implemented in this paper which can keep track of the patient's progress and sound an alarm in the event of danger. The MeanShift algorithm for real-time object tracking determines the location of the injection site. An area based detection method is presented to check whether the injection site is bleeding. Experiment results show that the video-tracking-and-monitoring system could track the injection site effectively and achieve a high tracking accuracy.