Research On Underwater Target Detection Using Multi-sensor

Marine target detection plays an important role in monitoring maritime safety.This paper based on a micro-buoy,achieve environmental information collection and passive target detection.The whole identification algorithm consists of three parts,noise signal processing,feature extraction and target classification.This paper focus on the following four aspects.Firstly,according to system function requirements,the overall structure of the buoy is designed.Then the hardware platform of data acquisition subsystem is designed and implemented,including the design of location and AD,SD card module with the sensor front end circuit.And then the PCB is designed and implemented.Secondly,in addition to the initial configuration of each module,the software design of data acquisition subsystem also contain storing and processing data.According to GPS data,MPU6050 and AD conversion,the following results canbe obtained: the attitude angle of the magnetic sensor,energy detection,and the azimuth power spectrum,LOFAR spectrum of the hydrophone etc.Thirdly,the underwater target detection model is constructed.The whole model consists of three parts,noise signal processing,feature extraction and target classification.The two adaptive denoising algorithms LMS and RLS are compared and analyzed.The Relief algorithm is used for feature selection to reduce the dimensionality of sample.The LVQ2 neural network classification using the optimal sample subset for training is performed.Finally,the paper gives the results of system test and simulation verification.The results show that all modules of the system meet the design requirements and can realize the collection,storage and processing of information.The classification feature is used to classify and identify the target features.The simulation results show that the marine target detection and recognition algorithm designed in this paper achieves the goal of predetermined design and has certain application value.