Adaptive Polarimetric Detection Of Sea-Surface Weak Targets Under The Complex Sea Conditions

Sea-surface weak targets detection is not only an important part of ocean cognition and ocean utilization,but also a solid foundation of the strategy of “Building An Ocean Power”.With the increasingly widespread application of fully polarized radar in sea-surface weak targets detection,fully mining the polarization information of the echo data and integrating the multi-polarized channels to design detection methods are the cruxes of improving the detection performance of sea-surface weak targets.However,the existing polarimetric detection methods have problems such as data model mismatch,which cannot adaptively match complex sea conditions,and strictly relying on the prior statistical information of sea clutter data,so it is difficult to effectively improve detection performance.In response to this problem,this thesis is devoted to the study of polarimetric detection method of seasurface weak targets with parameters that can be updated in real time under complex sea conditions.In the same time,it can eliminate the dependence on prior information of sea clutter and achieve adaptive detection of sea-surface weak targets.Firstly,a polarimetric detection method for sea-surface weak targets based on single texture factor is proposed in the case where the sea clutter textures of different polarization modes are less different.Specifically,the multi-polarized channels and correlations between different polarization modes are combined to construct a sea-surface polarized echo model based on compound Gaussian distribution in the multi-dimensional complex domain.On this basis,parameters such as the single texture factor to be estimated are introduced into the echo model to match different sea conditions,and multi-parameter joint estimation problems are established based on generalized likelihood ratio.Then,an iterative algorithm that can efficiently estimate the single texture factor and other parameters is proposed by using block-minimization majorization algorithm framework,which has linear complexity while ensuring convergence.Finally,the test results of the measured radar data sets verify the theoretical analysis proposed in this thesis.Compared with existing single-polarization detectors,the detection performance of the proposed polarimetric detection method is improved by an average of 4% and a maximum of 18.7 % in the same conditions.Furthermore,an effective multi-texture factors-based polarimetric detection method for sea-surface weak targets is proposed when the sea clutter texture difference between different polarization modes increases significantly under more complex sea conditions.Firstly,multiple texture factors are used to match the differentiated textures based on the polarized echo model of complex sea conditions,and multi-parameter joint estimation problems including multiple texture factors are established according to generalized likelihood ratio.Then,a block iterative estimation algorithm with linear complexity is proposed to solve the complex nonlinear estimation problem,which adaptively adjusts the parameters of the echo model according to the sea-surface echo samples,so as to obtain dynamic sea conditions in real time within effective convergence,thus ensuring adaptive matching of detection method and sea conditions.Finally,the test results of the measured radar data sets show that,compared with the single texture factor polarimetric detection method and the existing polarimetric feature-based detection method,the average detection performance of the proposed multiple texture factors based polarimetric detection method is improved by 12.16% and 8.53% respectively under the same conditions.In conclusion,this thesis devotes to the development and utilization of polarization information,and then designing high-performance sea-conditions-adaptive polarimetric detection methods,so as to achieve precise detection of sea-surface weak targets in complex detection environments.