No abstract

An efficient method for spaceborne synthetic aperture radar (SAR) geolocation is devised in this study. In this method, the pixels in the SAR image are divided into two types by a two-dimensional (2D) grid, that is, grid pixels and non-grid pixels. The approach of finding the position of a non-grid pixel is to add the position of the adjacent grid pixel and the position increment between the two pixels together, where the position increment is computed by means of three groups of increment formulae derived according to the principle of first-order Taylor approximation. It is noticed that this method is efficient and helpful in improving the efficiency of the geometrical SAR image registration for avoiding the polynomial fitting and interpolation. In the geometrical registration, the 2D shifts of a non-grid pixel can be expressed as the sum of shifts of the adjacent grid pixel and shift increments between the two pixels. Here the position and 2D shifts of the grid pixel are obtained through the numerical methods for a complete lack of closed-form solutions. Finally, the experiments with TerraSAR-X images are conducted to evaluate the effectiveness of the proposed method.

Due to the heavy computational complexity, the efficiency of FFBP algorithm descends obviously. Thus, an approximate FFBP algorithm is proposed in this paper. With the imaging plane on the sin r coordinate system, there is no need to calculate the arcsines of the back-projected angle. Meanwhile, the polynomial estimation method is introduced to obtain the approximate values of back-projected slant range and angle. As a result, the number of operations is further reduced. A detailed analysis of computation burden is given as well. At the end of the paper, dealing with the data of different wavebands, the simulation result of the point targets shows that the algorithm proposed is accurately and effective to give well focused images.