Research On SAR Image Basic Quality Evaluation

Due to the unique imaging characteristics of Synthetic aperture radar(SAR),the multiplier noise can be made by it.It is important that the evaluation of SAR images quality because of they have been used in many fields widely.Therefore,this thesis makes in-depth study for the quality evaluation method of the SAR images.Then,the main contents are as follows.SAR image quality evaluation method based on texture feature and structural similarity is proposed in this thesis.The number of the looks has many effects on the imaging results of SAR images.When the number of the looks is less,the SAR image noise is much higher,on the contrary,it is making the image easier to distinguish when the increase of the number of the looks for a certain area will be beneficial to the noise reduction of SAR images.Based on the relationship of the texture feature and the number of the looks in SAR images,this paper uses the gray level co-occurrence matrix to extract the texture feature and improves the similarity of the structure.Then,the SAR image quality is evaluated by the reference image quality and the concrete method is described in this paper.Therefor,we use the SAR images with different numbers of the looks to measure the quality parameters of the specific images in this paper,and finally verify the merits of the image quality.Then,the effectiveness of the method is proved by the experiments.In this thesis,a SAR image quality evaluation method with no-reference image is proposed,which is based on contrast and improved human visual system.The contrast is one of the important features that affect image clarity for both optical images and remote sensing images.The higher the contrast,the clearer the image will be,and the better the image quality.For SAR images,the magnitude of the contrast affects the clarity of the SAR image.However,the actual SAR image is generated without a reference image during the imaging process,which increases the difficulty of evaluating the original SAR image.Based on the contrast feature and the radar attributes of the SAR image,the human visual system is improved to evaluate the quality of original SAR image without reference image.Then,the contrast grade that it is the ratio which is combined with the different contrast and the different values of improving human visual system is obtained to evaluate the SAR image quality.In this paper,the quality of different original SAR images is evaluated.The quality of the original SAR images in the absence of reference images is evaluated by measuring the contrast grade of the specific images in no-reference images.In this thesis,an evaluation method of SAR image application ability for vehicle detection task is proposed.In practice,whether a SAR image is beneficial to an application is important for SAR image interpretation.This paper presents a new perspective that can assess whether a SAR image can be applied to vehicle detection.According to the basic characteristics of SAR images,the fitness parameters of SAR images are calibrated by using the request value and tolerant value of SAR images.The fitness of SAR images can be used to evaluate the SAR system,which the SAR image containing a vehicle is used for vehicle detection.