Research Of Microwave Stroke Detection Algorithm Based On FDTD

Recently, the incidence and mortality of stroke is increasing year by year. It is presented that the stroke has become the first cause of death of residents in our country by the statistic, which cause serious damage to people’s life and health. Medical studies have demonstrated that stroke can be timely detected could greatly increase the chance to cure and reduce damage of the brain. Stroke microwave detection is a kind of examination method which is to detect brain stroke by using microwave technology. Compared to the conventional detection means, Microwave detection has lots of potential advantages with security, excellent instantaneity and low cost, So it gets more and more attention of the researchers.Based on the research of the principle of microwave detection and brain dielectric structure, this paper firstly designed overall structure of the stroke of microwave detection system, which mainly includes the microwave signal source, the launching and receiving of antennas, microwave signal and microwave signal analysis processing module. Then simulation system establish the brain models by using the MRI brain images, deploy of antenna array around the brain, and generate the excitation signal using the modulating Gaussian signal. Then the stroke detection algorithm model is designed, which is composed of forward algorithm and inverse reconstruction algorithm. The forward algorithm use Finite-Difference Time-Domain method(FDTD) with perfectly matched layer(PML) absorbing boundary to simulate and calculate the brain region electromagnetic field. Aimed at the challenges, such as the irregularity, complicated structure(The forward algorithm do some electromagnetic simulation and calculation for the brain region using Finite-Difference Time-Domain method(FDTD) by absorbing boundary with perfectly matched layer(PML)). At the same time, aiming at the challenges, small features with dielectric properties differences and complex construction, reverse reconstruction designed a detection algorithm based on signal similarity using binary search iterative method, which can detect the clot in the brain and judge preliminary the existence and the direction. It is a modified particle swarm algorithm, can detecting the accurate information of the brain stroke clots’ size and location. Moreover, because of the accumulating stroke template database, the algorithm reduced the time of the reconstruction algorithm effectively.The brain model in simulation system is built by the MRI brain image which is around with the antenna array, and the excitation signal is modulated by Gauss. Research results show that the signal similarity algorithm based on FDTD can effectively detect the stroke clot; and the stroke detection algorithm based on the particle swarm can get the accurate information of the clots’ size and location. Finally, we optimize the PSO location detection system by adding the template signal database and an iteration termination condition, the efficiency is improved obviously. And the statistics show that the test time get more than 85% reducing.