【正文】 外文文獻：Extended Chirp Scaling Algorithm for Air and Spaceborne SAR Data Processing in Strip map and Scan SAR Imaging Modes Abstract This paper presents a generalized formulation of the extended chirp scaling (ECS) approach for high precision processing of air and spaceborne SAR data. Based on the original chirp scaling function, the ECS algorithm incorporates a new azimuth scaling function and a sub aperture approach,which allow an effective phasepreserving processing of ScanSAR data without interpolation for azimuth geometric azimuth scaling can also be used for automatic azimuth coregistration of interferometric image pairs which are acquired with different sampling distances. Additionally, a novel range scaling formulation is proposed for automatic range coregistration of interferometric image pairs or for improved robustness for the processing of highly squinted data. Several simulation and processing results of air and spaceborne SAR data are presented to demonstrate the validity of the proposed algorithms.I. INTRODUCTIONSAR IMAGE FORMATION is based on a coherent processing approach to build a long azimuth synthetic aperlure. This coherent processing can also be interpreted as the pression of a frequency modulated pulse, whereby the frequency modulation is caused by the natural movement of the sensor (Doppler effect). In addition, the transmitted pulses are also time dispersed and frequency modulated. This allows more energy to be transmitted (due to the time dispersion) and also a better range resolution to be achieved (due to the higher signal bandwidth). The range pressed pulse is obtained by means of a correlation of the received echo with the plexconjugated timeinverted replica of the transmitted pulse. Due to the curvature and the range variance of the azimuth modulation, SAR image formation is inherently a twodimensional (2D) process [7]. Commonly, the 2D processing is split into two ID steps in order to simplify the image formation process. This requires however an interpolation to pensate the range migration before azimuth processing. Another effect that occurs for a great amount of range migration (., for high squint angles and/or high range resolution or low frequency SAR systems) is the coupling of the range and azimuth signals, which is also range dependent. This coupling if not pensated leads to a defocusing of the range impulse response function (IRF). The distortion can be approximately pensated by means of a so called secondary range pression (SRC) which consists basically of slightly modifying the range modulation rate during range pressionThe first step in the case of the rangeDoppler processing approach [1], [5], [14], [25] is the range pression in the frequency domain. The SRC applied is correct only for one reference range and one azimuth frequency value (normally the Doppler centroid). The interpolation of the data is carried out in the rangeDoppler domain using a truncated interpolation applying the azimuth pression phase in this domain, the final image is obtained by an inverse Fast Fourier Transform (IFFT). The main disadvantages of this algorithm are the limited SRC and the need of an interpolation. The wave number algorithms [1], [3], [10], [1l] can be interpreted as 2D frequency correlations, whereby a 2D FFT is used to transform the signal from the time domain into the wave number domain. In these algorithms, approximations and the need for the so called Stolt interpolation increase heir implementation plexity and can degrade their phase preserving properties The chirp scaling (CS) algorithm allows the high precisionSAR processing without using interpolation in the SAR processing chain [21], [22]. It consists basically of multiplying the SAR data in the rangeDoppler domain with a quadratic phase function (chirp scaling) in order to equalize the range cell migration to a reference range, followed by a range pression and SRC in the wave number domain. Although the SRC is strictly correct only for one reference range, it is updated as a function of the azimuth frequency. The processing proceeds with phase multiplies and FFT operations, which make the algorithm extremely efficient. Reference [4] gives a detailed parison of wave number domain and chirp scaling processors. The extended chirp scaling (ECS) was developed originally for processing airborne data with strong motion errors (.,ESAR system, [12]) and with variable Doppler centroid in range and/or azimuth [13], [17]. The ECS algorithm allows the following steps to be included in the processing:Doppler centroid update with range by means of an azimuth spectral length extension in the rangeDoppler centroid update with azimuth by means of azimuth subaperture processing.Motion error correction for airborne processing by meansof an additional transformation into the signal domain. Improved processing of highly squinted data by means of a subtraction of the offset value from the chirp scaling function。First, SAR and ScanSAR systems will be briefly described. Synthetic Aperture Radar (SAR) is used for remote sensing and is increasingly employed for imaging, surveillance and exploration of the Earth39。s surface. This type of system uses a vehicle, such as an aircraft, helicopter, etc., traveling at a constant speed。 an antenna looking in a direction orthogonal to the direction of travel。 and a coherent radar system that periodically transmits electromagnetic pulses. The vehicle39。s direction of travel is termed the azimuth direction and the direction orthogonal thereto and slanted downward is termed range direction. In the course of a flight over an area to be observed, a swath, the length of the scanned stretch is imaged. The width of the swath depends, among other things, on the size of the time window in which the deflected radar echoes are received from a transmitted pulse. The received echoes are then frequencyconverted (mixed),