The Study On Target Detection Technique In Complex Sea Area

Underwater targets in complex sea area like ship,submarine and frogman will inevitably radiate noises to surrounded environment.These noises tend to contain significant target feature information,which is the foundation of a series of future work such as sonar detection,target identification and parameter estimation and is vital to our country's security.However,marine environment is variable,which makes it a really complicated wireless channel.In reality,radiated noises are prone to sink into marine environment noises;and as the technology of vibration damping and noise reduction become advanced,small radiated noises is harder cached and detected.Therefore,this paper thoroughly studies the problem of multi-pattern targets detection in complex sea area.This paper,starting with marine noise environment,analyzes the physical transmission characteristics of marine sound channel,and builds a model of mirror double extension and impact noise channel and an underwater targets radiation noise model of ships and frogmen,etc.With the guide of chaos theory and stochastic resonance theory and combined with the real requirement of marine engineering,this paper put forward some new methods as follows:First of all,aiming at the strong noise of complex sea area and at the nonlinear character of water-sound channel,the Maximum Correlated Kurtosis Deconvolution(MCKD)detection method is proposed.This method initially detects target signal in complex sea area in real time through an m-order MCKD.when an invaded target(such as ship,frogman)is detected,MCKD will output the maximum peak and pulse response of target noise signals,and extract signal clips of pulse response,further,with the help of chaos system,frequency,phase and amplitude of target signal can be detected.Meanwhile,in order to resolve the problem of fuzzy judgment for chaotic critical and large scale period state in chaotic system,the Menikov Function and criterion for zero-crossing detection is employed and a small signal detection system is built based on chaos oscillators array.Secondly,in order to cope with the problem of high-frequency small signals being hard to detect by stochastic resonance system and the problem of multi-frequency small signals,this paper proposes the method of heterodyne and sweep-frequency stochastic resonance,and takes four-propellers-ship as the object to verify this method.The results show that the optimized system has an obvious increase in output SNR.Ultimately,this paper analyzes and designs a multi-pattern targets detection system,and carries out a simulation experiment with real underwater radiated noises data of ship,submarine and frogman.Simulation results indicate that the ship-radiated noise can still be detected in the condition of low SNR,which verifies the algorithm demonstrated in this paper,and thus expected study goal is accomplished.