Research On Key Technologies And Applications Of Net Dome Camera

Integrative Dome Camera is a highly integrated, device which is combined with optics, mechanical, electronic, image and information technology. It is a uniting of intelligent machine and image information technology, and it is of great importance to surveillance area. Net Dome Camera is a significant extension for this technology, which solves the shortcoming in traditional surveillance that treats PC support as a must. Since it has OS embedded in its body, it can have image collected and transferred independently. Besides, Net Dome Camera use networks replacing cables in data transition, which not only simplifies the system, but also extends the area of surveillance. Based on the Integrative Dome Camera that is independently developed by our research group and also formally produced by now, this thesis performs in-depth researches and experiments on some key technologies of Net Dome Camera, which is mainly applied to WAN service.In this work, the theory and development of surveillance based on image is discussed firstly, as well as the deficiency in traditional surveillance system and its improvement orientation. Then this thesis reviews the general situation of Net Dome Camera, and the main content of the thesis is presented.The constitution theory of Net Dome Camera is brought forward afterwards, and the key technologies related are fully analyzed from aspects including Integrative Dome Camera, embedded video server and client software.Considering of the point that Net Dome Camera is running on the internet, the thesis put its emphasis in doing studies on network technology. In this work, RTP / RTCP protocol is chosen as the basic protocol for image transition, which ensures the speediness and real time of video. Besides, a control theory for output bit rate and encoding rate is brought forward, which solves the problems caused by net crowding. Then a result analysis is presented.Object detection and recognition is a key technology for Net Dome Camera, which imitates the human recognition. Therefore, object detection and recognition is studied in this work as well. Based on the study, this thesis proposes a method for object detection based on color information. Besides, a method for status recognition based on Fuzzy Hypersphere Neural Network is presented as well. This work also applies the study to the fault detection in pneumatic system. Experiments show that the ratio of successful recognition is 99%.Based on all the research results, a sample system of Net Dome Camera for WAN is developed, which has the functions including video collection, coding, transition, remote movement control and client software for decoding and playing.Finally, the last past gives the summary and conclusion for the whole thesis, points some advice for improvement and forecasts the trends for future.