The proposed paper deals with an automatic trajectory monitoring system designed for the rendezvous between the automatic transfer vehicle (ATV) and the Iautononternational Space Station (ISS). During final approach phase, a TV camera on the ISS currently provides images of ATV visual targets to be used by ISS crew for visual monitoring. The proposed monitoring system, based on a dedicated image-processing algorithm applied to these TV images of the approach, is intended to autonomously and automatically determine relative ATV-ISS position and attitude. Artificial intelligence techniques for edge detection, Hough transform and pattern matching are used to implement a recognition algorithm, able to fast and accurately indicate ATV visual targets position and orientation with respect to ISS. Those values are then processed in order to calculate the full set of relative ATV navigation data. Finally, an estimation of the relative trajectory evolution via a classical astrodynamic tool as the Euler-Hill relative motion equations follows, enabling the crew and the control centre to verify that trajectory corridors and attitude constraints are respected. According to ATV mission constraints, which require using existing sensors (i.e. ATV visual cues and ISS TV camera), the performances and the accuracy of the monitoring system are evaluated for significant approach cases. Overall, computation time and hardware requirements of the proposed system easily fall into the limits foreseen for a real-time on-board application. Both the ATV Control Centre (real-time GNC data check against telemetry) and the ISS (automatic monitoring tool for ISS crew) could take advantages of the output data provided by this system. Moreover, obtained results clearly show that the monitoring system is a suitable candidate for further developments and improvements aimed to provide a complete support tool for ATV mission operations during final approach phase.