Characterizing A Target And Its Applications Based On Polarimetric SAR

A traditional radar provides amplitude information or RCS of a target, but some information on phase is lost. However, a polarimetric radar provides a 2 by 2 complex matrix of a target, not only amplitude but also phase. Therefore, a polarimetric radar can obtain times of information than a traditional radar. Because of difficulties in making a polarimetric radar system, there is no progress in the field of radar polarimetry for a long time. As the sophisticated theory coming and electronic technology developing, radar polarimetry has become a indispensable tool, and polarimetric radars have been applied in several fields such as military, remote sensing, agriculture, hydrology and monitor, and so on. Characterizing a target and its applications are investigated in this thesis. Our works focus on how to characterize a target and process a SAR image, and some results are obtained as followings:A method on extracting the characteristics of a target based on eigenvalues of a scattering matrix is proposed. This method eliminates the effects of target orientation angels and power of radar echo. From the derived parameters, we can analyze the plane and diplane components of a target. A SAR image collected in Niigata Japan was studied with our method. According the results, we can say that the method is effective in characterizing a target.We enhance the contrast between targets and background with a simple combination of polarimetric parameters, including total span of a scattering matrix, polarmetric entropy, eigenvalue similarity parameters. Meanwhile, we bring the method of image balance into SAR image, so the distribution of brightness in the image is renewed. For the mechanism of imaging radar, it is inevitable that the coherence speckle advents. We also bring the method of median filtering into image process in order to eliminate the impacts of coherence speckle noise. It is helpful for the further processing image well after above processes.A universal method is proposed for extracting the characteristics of roads from an image. Up to now, traditional methods employ template matching with different operators such as Canny Operator, Sobel Operator, Log Operator. Because of the speckle noise, traditional methods do not work well for some cases. So we present a new method to extract road in an image automatically. The new method has the features of larger neighborhoods and more precision with relatively less calculation. In the new method, we also use an iteration algorithm in order to reduce noise for keeping the continuity of a road.