Abstract
In recent years, wireless vision sensor network (WVSN) is being used to retrieve video content from different image sensors which are connected to different devices wirelessly. This information is used to do video analysis which can help to automate different tasks such as video surveillance. For such systems, power consumption during processing and communicating information has been a challenge because of limited energy sources at the node. To deal with the energy consumption problem, in this paper WVSN is proposed with its algorithm and hardware implementation for a smart home application. The computation tasks have been divided between the sensor node and the central server. While taking care of privacy issues during the transmission of data, a low complexity system has been developed for sensor nodes. For video analysis, foreground segmentation, object labeling, and object tracking have been performed. An efficient binary data compression technique has been proposed to compress the information during the labeling process. The proposed system has a high recognition rate for gesture recognition and human tracking. The system can achieve eight frames per second during processing information on Raspberry Pi board. This system can be extended further to include other smart home applications.
Highlights
Smart home is become a mainstream trend in products during recent years due to a lot of advancements to facilitate human life
In the second type of strategy, image processing is performed at vision sensor network (VSN) and feature results, which are acquired after image processing algorithms, are transferred to the server for further analysis
The whole system has been implemented on an embedded board as a demonstration system
Summary
Smart home is become a mainstream trend in products during recent years due to a lot of advancements to facilitate human life. In the second type of strategy, image processing is performed at VSN and feature results, which are acquired after image processing algorithms, are transferred to the server for further analysis. Examples of such systems include DSPcam and CMUcam3 [5], [7]. The whole system has been implemented on a well-known WSN platform, Raspberry Pi. The main contributions of the paper include the following: 1) The combination of computer vision processes and image compression: the computer vision part contains the operations of segmentation, labeling, and tracking.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
