| Last century has seen the rapid growth of elderly population throughout the world. Many countries, including China, have entered the aging society. It has caused international concern to cope well with this issue while satisfying the needs of elderly population. Ranking top among major causes of elderly injuries, the issue of accidental fall has drawn attention from researchers both at home and abroad. Timely and effective recognition of such falls will minimize the respond time, as well as injuries of casualties. In this regard, the care service of elderly people is of great significance.Traditional medical care system has already realized the model of local collection and local processing by relying on X86 processor or microprocessor as hardware platform, with software programs for detecting video images from cameras. While the FPGA system with cloud computing features is able to process data collected by remote devices, which meets the demands of users for huge amounts of signal requests and remote collection. Realizing remote collection and cloud processing, the FPGA system is of powerful processing ability and higher efficiency.This research designs a cloud care system based on PCI express, facing local or remote video images on Xilinx Virtex 6 FPGA platform, which can be applied in indoor space with good visibility. The FIFO streaming microarchitecture in the FPGA system can realize the real-time fall detection of each images collected by local and remote cameras. The PCI express transmission channel in the system connects the FPGA with a PC server, supplying video frames received by PC for subsequence detection. The fall detection module in FPGA chooses a modified algorithm that is more suitable for the streaming microarchitecture. Compared with traditional implementation by software, the FPGA system in the article enjoys a lower latency, a higher speed and a better throughput, and it may be extended to a general medical nursing system in further research. |
PDF Full Text Request