THE SYSTEM AND THE METHOD FOR MONITORING ILLEGAL STOPPING AND PARKING VEHICLES BASED SPATIO-TEMPORAL FEATURE VECTOR-BASED VEHICLE RECOGNITION

摘要

The present invention relates to a system for curbing illegal parking and stopping vehicles using spatio-temporal feature vector-based vehicle recognition, which is capable of efficiently curbing illegal parking and stopping vehicles by recognizing and controlling vehicles in images taken by a detecting camera using spatio-temporal feature vectors, and a method for the same. A system for curbing illegal parking and stopping vehicles according to an embodiment of the present invention comprises: a crackdown camera; a detecting camera for taking a detecting image for a parking and stopping prohibited zone; a memory for storing learning data including feature vector information for a vehicle image; and a control unit for extracting a first feature vector from the detecting image using spatio-temporal feature vectors, and comparing the first feature vector with the learning data to recognize vehicles among objects included in the detecting image, wherein if a first condition associated with a first vehicle which is a part of the recognized vehicles is satisfied, the control unit controls the crackdown camera to perform at least one among panning, tilting, and zooming to take an image of the license plate of the first vehicle, wherein the first condition may be that the first vehicle stops more than a preset time in the parking and stopping prohibited zone. [Reference numerals] (AA) Start; (BB) End; (S1010) Photograph a first detecting image for a parking and stopping prohibited zone; (S1020) Extract a first feature vector from the first detecting image; (S1030) Recognize vehicles from the first detecting image using the first feature vector and learning data; (S1040) Is an image size of a vehicle recognized from the first detecting image within a range of a size calculated in response to a first distance?; (S1050) Wait more than a preset time; (S1060) Photograph a second detecting image for the parking and stopping prohibited zone; (S1070) Extract a second feature vector from the second detecting image; (S1080) Recognize vehicles from the second detecting image using the second feature vector and learning data; (S1090) Are the vehicles recognized from the first detecting image and the vehicles recognized from the second detecting image matched?; (S1100) Curb the matched vehicles using a crackdown camera; (S1110) Add feature vector information for the vehicles to the learning data