Pretreatment Of Image Based On X-band Marine Radar

Sea wave is such a complex phenomenon that the research on it makes great significance on ocean engineering construction and development,energy detection,ship transportation,fishing and so on.The monitor of sea wave begins in the early 1950 s,and the way to get value by X-band radar image becomes hot in recent twenty years.On the basis of a studying project,significant wave height,wave direction and wave period of the waves can be obtained through a series of calculations on the collected radar echo images.Then,the waves can be real-time monitoring.In static environment,wave related parameters can be accurately estimated.That is to say,waves monitoring has achieved remarkable results in static environment.But in dynamic environment,the wave power spectrum occurs frequency shift since the speed of the ship is too large.In this case,radar speed is not enough which will produce aliasing problems.Therefore,this paper focuses on the analysis of dealiasing problem,and proposes the corresponding solutions.First of all,in dynamic environment,because of the Doppler effect,the waves frequency will increase on the direction of faceing the ship motion,and will decrease on the direction away from ship motion.In this case,the sampling frequency of the radar can not meet the requirements for the sampling frequency,the aliasing will produce.Eventually,the parameters estimation of waves will be affected.However,the compressive sensing theory can solve the problem of insufficient sampling frequency.Therefore,this paper intends to use a common method of compressive sensing-orthogonal matching pursuit method(OMP),which can achieve dealiasing of radar images.By the numerical simulation,proving that this method can effectively estimate the parameters of the waves.Secondly,on the basis of the theory of compressed sensing,this paper proposes a signal reconstruction algorithm based on the Discrete Particle Swarm Optimization(PSO).Currently,the more mature reconstruction algorithms of compressed sensing are transforming the NP-hard problem of 0 norm into other optimization problems of same solution.Such methods based on the theory of numerical calculation.By introducing a parameter ?,an optimization problem with constraints transforms into an unconstrained optimization problem.Then,the optimization problem can be directly solve by using the discrete particle swarm algorithm,this method is simple and easy to understand.Finally,the effectiveness of this method is proved by the numerical simulation.This method can reconstruct signal and image,and better than the OMP algorithm,it provides a new way for reconstruction algorithm of signal and image.