| 摘要: | 研究期間：10205~10304;In recent years, Wireless network provide interconnection between electronic devices without cable or wires. Technologies such as CDMA2000 、GSM、WiFi、 WiMax were applied on many applications .With rapid i mprovement in technology, living environment becomes different, a new concept, “intelligent life”, which substituting the traditional environment, is developed and accepted in our lives. By Combing the devices , we can control them via PC、PDA or cell phones , live becomes much more convenient and different . In the past few years, Wireless sensor network was used to only sense temperature, moisture, vibration and so on. However, we add the concept of “vision.” In this project, the “Joint Source and Vision Sensor Network ”scheme is introduced. Users give an instruction at fronted interface, and then the control panel will drive the devices via Wireless sensor network to achieve human -computer Interaction. To implement the object recognition system, people u sed to acquire video image after compression, then execute Foreground Segmentation. And then achieve Event Detection function. But the shortcoming is loss of privacy. It cannot protect personal information. Therefore the foreground segmentation plays an im portant role in pre-processing in our project. Because of the limitation in small -scale camera and the privacy in wireless transmission, not all the means can apply on such restrain. In the field of image and video processing, object segmentation has bee n developed in many years. However, most of the methods have high complexity and loss of privacy; we proposed a segmentation technique in small -scale camera to transport in binary arrays. Comparing with whole image, transporting an image of object has much less information and binary information cannot revert to original one. Therefore, we reduce the amount of information and achieve the protection of privacy .Then, we implement human body detection or gesture detection with those data we processed. Human body detection can apply on anti -theft systems, safety monitoring systems, control the direction of blowing in air conditioning, etc. On the other hand, gesture recognition can detect the behavior of human, such as controlling of mouse, electronic devices, light, widows, etc. |
