Underwater Target Localization And Inversion Technology Based On Magnetic Anomaly Detection

Due to the concealment of underwater environment,traditional acoustic,optical,and electric detection methods all have their limitations and cannot meet the requirements of efficient detection and positioning of underwater targets.Most of the underwater targets are ferromagnetic targets,which are affected by the magnetization of the geomagnetic field.So that the magnetic field distribution in the surrounding space will change,thus producing magnetic abnormal signals.Location and inversion of underwater target based on magnetic anomaly detection aims to obtain characteristic signals related to target position and distribution,by using the spatial distribution of magnetic anomaly signals generated by underwater ferromagnetic target,to realize the detection,localization,and recognition of underwater targets.In this dissertation,according to the requirement of different underwater target detection tasks,underwater target positioning based on static and dynamic magnetic detection platforms,and underwater target inversion technology with multi-unmanned aerial vehicle cooperation are studied.In the field of magnetic anomaly detection,there are two main problems.One is that,since both the target and the detection platform are exposed to the geomagnetic environment,the target signal contains strong geomagnetic background magnetic field and the platform interference magnetic field,so that it is difficult to achieve effective separation of the target signal.Second,because the magnitude of the magnetic anomaly signal attenuates in a cubic trend with the target distance,the signal-noise-ratio in remote detection is low.Therefore,in view of the above problems,the main work of this dissertation includes:First,in the field of static platform magnetic anomaly detection such as shore-based underwater magnetic anomaly target warning,the environmental magnetic field changes in real time,and the detection task requires real-time estimation of the target position,so high requirements are put forward for the detection platform and detection algorithm.Since magnetic anomaly difference signals can effectively avoid interference of environmental magnetic field changes,the characteristics of magnetic anomaly high order quantities have been studied in the dissertation.Furthermore,the magnetic gradient measurement framework and real-time analytic positioning method for underwater targets with unknown magnetic moment and known magnetic moment are proposed,which effectively improve the positioning accuracy and robustness of the static platform.Second,in the field of the dynamic platform magnetic anomaly detection such as the anti-submarine task,traditional antisubmarine aircraft carries scalar magnetometer.Since the scalar magnetic field takes inadequate information,and it is vulnerable to background magnetic field and interference magnetic field of the carrying platform,coarse positioning of underwater target can only be done by multiple reciprocating flying,which cannot meet the military requirements of the modern marine warfare.In the dissertation,through the analysis of the high-order magnetic characteristic of the target in the dynamic detection process,the orthogonal basis decomposition and correlation detection method based on the magnetic gradient measurement are proposed,and the frequency characteristic of the target signal is analyzed,which enables the signal-to-noise ratio further improved by low-pass filtering.The method can effectively reduce the magnetic interference of the environment and the carrier,furthermore,by using the characteristic time search algorithm to determine the vertex distance at one time,the target real-time localization in the process of dynamic magnetic anomaly detection has been achieved.Thirdly,in order to realize the characteristics recognition of underwater magnetic anomaly target,it is necessary to establish the simplest equivalent model of different types of target.In view of the problems that the magnet simulation number is too many,and the equivalent magnetic dipole magnetic moment direction is not clear,which cause the difficulty of solving the equivalent equation.In the dissertation,the influence of the magnetic moment and spatial distribution of multi-magnetic dipoles on the equivalent effect with different spatial scales is studied,that is,the equivalence law of multi-magnetic dipole equivalence.And the simplest equivalent magnetic dipole model is established.The multimagnetic dipole equivalent model of different types of targets is used to generate training samples of high-order magnetic distribution at multi-spatial scales.By using the principle of deep convolutional neural network in image classification and recognition and border regression,the target classification and recognition of underwater targets with UAV group drawing multi-spatial scale high-order magnetic distribution map is realized.Fourth,a multi-channel synchronous magnetic signal acquisition device is designed,and a static compensation method for magnetic gradient measurement system based on genetic algorithm and a dynamic carrier magnetic compensation method based on vector measurement are proposed.The detection and positioning tests of equivalent magnetic anomaly targets under dynamic and static platforms are carried out respectively,and the test results verify the real-time positioning ability of the algorithms.