With the advent of CMOS cameras, it is now possible to make compact, cheap and low-power image sensors capable of on-board image processing. These embedded vision sensors provide a rich new sensing modality enabling new classes of wireless sensor networking applications. In order to build these applications, system designers need to overcome challenges associated with limited bandwith, limited power, group coordination and fusing of multiple camera views with various other sensory inputs. Real-time properties must be upheld if multiple vision sensors are to process data, communicate with each other and make a group decision before the measured environmental feature changes. The appearance of Wireless Multimedia Sensor Networks (WMSNs) requires a new generation technology of image processing for many applications. This paper presents a new approach to License Plate Localization(LPL) in WMSNs. We detect the license plate by a variational model that is embedded in several scalar functions. The motion of the dynamic interface is governed by nonlinear Partial Differential Equations(PDEs). Such variational models are flexible in handling complex environment and are concise in extracting plates boundaries despite of serious pollution. The cost of this method is moderate. The accuracy and efficiency of the proposed algorithm are illustrated by several numerical examples.