Algorithm Research On Ship Detection Based On SAR Image

In recent years,Synthetic Aperture Radar imaging technology and applications have been developed rapidly.Synthetic Aperture Radar provides an another important way for ship detection while it exists a detection blind spot in Navigation radar and AIS can not provide non-cooperative ship information.However,the detection problem under the complex background SAR image has not been solved very vell.In this paper,the detection method of ship target in SAR image with complex background is given,and a solution is proposed.The visual attention mechanism is introduced and combined with the Shearlet transform detection method.The experimental results show that the detection effect is good.The main works are as follows:The statistical distribution of sea clutter in complex background is studied.Several kinds of sea clutter statistical distribution models and its parameter estimation method are introduced.Then use the models to fit the sea clutter,and analyze the fitted results.The visual attention model is combined with CA-CFAR detection for pre-screening of ship targets.Select the SR visual attention model to calculate the significant area,generate a significant map.Do the CFAR detection on the significant map to obtain the initial screening results.The SR model is studied,and the average logarithmic amplitude spectrum of SAR image is analyzed,and the feasibility of applying the model to SAR image is clarified.The clutter distribution of the original image is changed after the significant area calculation,and the sea clutter intensity is concentrated to the low intensity region.The evaluation of the clutter intensity and the inhomogeneity is the mean value and standard deviation.The results show that the mean value and standard deviation of the clutter are significantly reduced.In this paper,a quadratic significant graph is proposed while much clutter still exists after CFAR detection on the first significant graph.The experimental results show that the clutter interference is excluded.For the false alarm phenomenon caused by the significant graph,the Shearlet transform detection algorithm is used to judge the false alarm.The Shearlet transform detection method is prone to false alarm at the edge of the strong clutter and has the limited image size.In this paper,it is used to identify the false alarm,thus the detection area is limited to a fixed region of the CFAR result,and at the same time an improved method is given for the complex background.After the coefficients are weakened or enhanced according to the adaptive threshold threshold,multiplying operation across the scales is carried out in each direction,and then the coefficients are fused at the direction.Through calculating the contrast parameter value of this improved method and the predecessor method,the results show that the improved method has a larger contrast value.