An automatic monitoring approach for unsupervised parking lots in outdoors

Abstract

A video-based monitoring approach for outdoor parking lots has been developed in this paper. Since parking lots located in outdoor were the open and widespread spaces, moving objects in grabbed images were too small to obtain the detail and recognized image for the object, identification, object behavior analysis, or illegal event alarming. A dual-camera device was designed and calibrated manually. Using the calibrated parameters, multiple target images with high quality can be grabbed from widespread open spaces. The tracking process for a specified target can be easily switched to another when multiple objects appear in the monitoring space. All of them are stored in video-based databases of DVR systems. In addition, the object images are processed for the furthermore retrieval. Some experimental results were given to show the validity of our proposed approaches