Research On Ship Target Detection Technology In SAR Image

Synthetic aperture radar is an active radar.Its working band is mainly the microwave band.It has a strong penetrating power to the space medium,and is less affected by factors such as light.Therefore,it guarantees its performance at night and in various weathers.Work efficiency,and this feature is of great significance for surface target monitoring.With the improvement of the resolution of the synthetic aperture radar,SAR images contain a wealth of information,including the shape of surface ships,the types of ships,the wakes of ships,etc.,and the corresponding parameter information.From a civilian level,this type of information can provide the location of the target and is of great significance for directing maritime traffic.From a military perspective,this type of information is important for our army’s perception of battlefield conditions and command operations.It can know each other and provide solutions in advance.Therefore,the interpretation and interpretation of SAR images has a great effect on safeguarding the rights and interests of the oceans.As an important part of SAR image interpretation and interpretation,it is of great significance to study high-efficiency and high-accuracy SAR image ship target detection methods.This paper first introduces the constant false alarm rate detection algorithm in the classic algorithm of ship target detection in SAR images,analyzes various constant false alarm rate detection algorithms and the marine background clutter model in the algorithm,and gives the algorithm’s Scope of application and the process of parameter estimation and calculation of relative thresholds for various marine background clutter models.This article mainly focuses on the modeling of the marine background clutter,and in order to better detect For near-shore ship targets,a detection method based on the ship’s geometric features is proposed in this paper.(1)A constant false alarm rate detection algorithm based on Gumbel’s mixed distribution model was proposed.Constant false alarm rate detection algorithm is one of the most widely used and best detection algorithms for SAR image ship target detection.It detects the target in the image based on the gray value characteristics of the SAR image,and uses the target gray level.The difference between the value and the background gray value is large to detect potential targets in the image.In this paper,a Gumbel hybrid distribution model is proposedto describe the background clutter of the ocean.The position and scale parameters in the Gumbel distribution model are used to better describe the background clutter of the ocean.At the same time,the position in the Gumbel mixed distribution model is given using the expectation maximization method.The parameter estimation process of parameters,scale parameters and weights,and the effectiveness of the model and the effectiveness of the parameter estimation method are verified by the effect of model fitting.Finally,the Newton iteration method is used to give the threshold estimation process of the constant false alarm detection algorithm based on the model.The simulation results of the measured SAR data are used to verify the superiority of the proposed model compared to other models.(2)An algorithm of near-shore ship target detection based on geometric features is proposed.For the detection of near-shore ship targets,as the resolution of SAR images gradually increases,the ability to describe ship targets has also been greatly enhanced,and more and more detailed information is contained.Based on this,this paper proposes ship-based targets.Detection method of geometric features.Aiming at false alarm targets,a method based on a priori knowledge of ship targets was proposed to remove false alarm targets.The algorithm first preprocesses the SAR image,then extracts the geometric features of the ship target,completes the ship target detection,removes false alarm targets at the same time,and gives the final detection result.Finally,the effectiveness and accuracy of the proposed algorithm are verified by simulation experiments on measured high-resolution SAR images.