| Coal is a non-renewable resource; it plays a vital role in the development of national economy. Especially under the new market mechanism conditions, coal faces fierce competitive market. Therefore, the exploitation and utilization of mineral resources should be monitored effectively, in order to achieve maximum protection.The system is proposed, based on the maximize protection and utilization of coal resource. Video surveillance system is designed and installed in the entrance of coal market in order to monitor vehicles and get the decision of giving green light or not. Target tracking and license plate recognition technique is combined in the system, in order to surveillance from a single point. By PTZ rotating, target image is scanned from different point; in order to obtain a clearer picture, improve accuracy and fast of the system recognition.In the whole process of target tracking, the continual frame distribution has been used to detect the moving target, the twin-peaks approach has been used to select the optimal threshold. The region of moving target is obtained. Parting off target and background, image binaryzation takes shape. Some isolated point and cavity exists in the view picture now. These unhitched sounds will have influence for the performance of the system. Morphology filtering is recommended to dispose these noises. After disposing the whole picture, predict the target location of next moment by using Kalman filtering, accomplishing trace. Pass the scanning to upper computer to distinguish. If it can be distinguished, the picture will be collected. Result will be sent to license plate recognition system to be checked, and the decision of giving green light or not.In this paper, selection connection and testing are achieved by hardware. Target tracking module and the license plate recognition module, PTZ control module are designed in Delphi7.0. During the simulation in the laboratory, the function of the system is reliable. |
PDF Full Text Request