Vehicle localization is an important component of intelligent transportation systems and telematics applications. Localization systems typically rely on Global Positioning System (GPS) technology; however, the accuracy and reliability of GPS are degraded in urban environments due to satellite visibility and multipath effects. We propose to use a Kalman filter to fuse data from a GPS receiver and a machine vision system to position the vehicle with respect to objects in its environment. Data association is needed to identify the detected objects, and to identify the road driven by the vehicle. For this purpose we employ multiple hypothesis tracking to consider multiple data association hypotheses simultaneously. Experimental results show that using machine vision reduces the effect that GPS measurement errors have on localization accuracy. Vision also improves the identification of the road being driven by the vehicle, which is important for the problem of map matching in vehicle localization.