A system-level approach for indoor surveillance of objects using wireless sensor networks and Pan/Tilt camera is presented in this paper. The objects to be tracked carry a listener cricket mote and beacon motes are mounted on the ceiling. The time difference of arrival (TDoA) approach and multilateration are used to estimate the distance and position of the object, respectively. The small error in the position estimation result was further reduced by implementing a Kalman filter. The position of the object obtained from the cricket motes are used to calculate the pan and tilt angles of the camera, also mounted on the ceiling. Since the camera used does not have the zoom capability, a laser pointer is used to constantly point a dot on the moving object. The pointer followed the object with negligible error.