A Study of Embedded Surveillance System for Home Security based on Image, Audio, and Vision Technologies
The goal of this research proposal is to develop a real-time video/audio surveillance system for home security, as well as to implement its integrated real-time embedded platform. We will firstly develop a fast hybrid DWT and DCT-based technique to accomplish the requirements for real-time video compression, including high speed, high image quality and high compression ratio. By means of the spatial and sequential characteristics of surveillance images, this technique applies a low-complexity and low-memory cost zero-tree entropy coder for DWT and DCT
coefficients to accomplish an effective technique for encoding and decoding surveillance videos, which will greatly improve the speed of compression and decompression, as well as maintains images of high vision quality. Thus, this planned technique will provide
satisfied quality and smoothness even under multi-channel video surveillance and embedded implementation. Besides, we will also develop a fast motion detection, target object segmentation and tracking approach, to provide a technique for optimizing the coding efficiency of the video codec, saving the storage cost, and provide warning information for possible invasion and danger. To provide further audio information, we will also implement a fast ADPCM- based audio coding technique for supporting the purpose of audio surveillance. Moreover, an RTP (Real-Time Transport Protocol) model for real-time transmitting the surveillance video/audio streaming via network will also be integrated to our newly developed system.
Finally, the above-mentioned surveillance techniques will be integrated and implemented on an ARM-DSP dual core embedded platform, as well as the computer peripheral systems, including image grabbing devices, storage devices, wireless network, GSM module, mobile communication devices, and other control
interfaces, will be also integrated to accomplish a versatile embedded video/audio surveillance system for home security.
Keywords ：Video/audio surveillance, real-time video compression, wavelet transform, DCT-based coding, zero-tree coding, remote mobile surveillance, embedded vision system
