The Design Of Target Tracking And Recognition System Based On The Android Wireless Sensor Network

The Wireless Sensor Network (WSN), one of the most active technological fields at present, is composed of small-sized sensor nodes that can collect, transmit and process data, and it has an extensive applying prospect in daily life, industrial production and the school research area. As the computer vision and Android applications develops, together with the progress of chips design and manufacturing technique, it is the certain trend to transplant the Android system into the embedded devices. Android system has a powerful interactive capabilities, excellent display effects and dazzling graphical interface, but the ability to self-image processing, target tracking is relatively limited. OpenCV--computer vision library contains a lot of methods, which efficiently accomplish this task, and are also supported by the Android system. It becomes possible to use OpenCV on Android.The moving object detection is the precondition of target tracking. At present, there are lots of target tracking algorithms in the field of image processing and computer vision research, such as the classic meanshift, centroid algorithm, and tracking algorithm based on template matching. These algorithms are reliable in the target recognition but barely satisfactory in the target tracking. TLD target tracking algorithm is composed of three parts, namely, tracker, learning process and detector. The strategy of combining tracking and detecting is an adaptive and reliable tracking technology.In this paper, the design of a wireless sensor network nodes as Android embedded devices to achieve the target tracking and recognition functions. Master chip is Samsung’s S3C6410, on which to build the Android OpenCV development kit, using TLD target tracking algorithm, so that the validity of the data collection node, interactivity have been improved. Data transmission between nodes are interconnected using Wi-Fi technology, saving costs, while for short-distance transmission reliability can be guaranteed. Finally, testing the device on target identification and tracking results prove the effectiveness of the system identification.