The Application Of Intelligent Video Surveillance In Custom Dock Intelligent Inspecting System

Intelligent video surveillance is an important research field in computer vision. It has promising prospect in the applications of collecting traffic information, and monitoring the important areas, such as bank, hotel, and railway station.In this thesis, we focus our research on some issues in detecting and tracking objects. As applying our algorithm we develop a practical Custom Dock Intelligent Inspecting System(CDIIS). The major contents can be listed as follows:1 . In aspect of moving object detection, after analyzing the advantages and disadvantages of several commonly used detection method, an effective scheme is proposed to improve the mixture Gaussian model algorithms for the detection of objects in complex and ever-changing environment. At the model initialization stage the proposed algorithm first gets an accurate background model; and then introduces accelerating factor and reasonable feedback at the model updating stage, which will thus acquire the background faster and more accurately without any losses of the sensitivity in detecting objects.2. A static vessel detection and extraction algorithm which is based on background model is proposed as required by Customs. Experimental results demonstrate that vessels can be detected and extracted accurately in different scenarios by using the proposed algorithm.3. By characterizing the object region with the region of object's external polygon, existing multi-target extraction algorithms based on the object marking for moving object extraction is improved to minimize the defection of object which arise at object detection stage, while keep the object region as accurately as possible.4. For object tracking, prediction and tracking based on Kalman filter and the wharf incident detection method are applied. The results have been achieved for a practical CDIIS.