The majority of today’s video surveillance systems are analog based, requiring human interaction and are susceptible to multiple threats. Utilizing a fully digitized system incorporating the Vector Bank (VB), Laplacian of Gaussian (LoG), and Directional Discrete Cosine Transform (DDCT) to each surveillance camera in a surveillance network, the need for digital processing would be accomplished at each individual camera level, reducing the memory required at the network hub and providing a decentralized secure surveillance network. Detection and tracking of moving objects may be sensed by each independent surveillance camera in the network, processed using VB, LoG, and DDCT to drastically reduce bandwidth, and computation time of processing, ultimately reducing data bus traffic and transmission bandwidth. In this paper, the addition of a hardware based System on Chip (SoC) digital signal processing algorithms interfaced to H.264 architecture for the purpose of a fully redundant and autonomous multi-camera digital video surveillance network is discussed. Experimental results show the performance and its capability to detect, isolate and track the objects, while reducing computational time, channel bandwidth and overhead costs of the digital video surveillance network while increasing security. Such achievements prove to greatly improve computer visualization, and performance of the network.