|
With the continuous advancement of constellation remote sensing technology for observing star clusters, in recent years, researchers both domestically and internationally have been increasingly focusing on the application of Synthetic Aperture Radar (SAR) observations from multiple stars. However, even with constellations having hundreds of satellites, individual calibration is still required for each satellite during radiometric calibration. Due to different imaging parameters of various satellites, this may lead to differences in the grayscale values of imaging results in overlapping areas, adding complexity to data fusion and applications. Therefore, there is currently a lack of a technical solution for evaluating the radiometric consistency of multi-star SAR. To simplify the complexity of radiometric calibration evaluation and address the challenge of assessing radiometric consistency due to differences in viewing angles in multistar data acquisition, this paper proposes a homogeneous region-based method for evaluating radiometric consistency among multiple stars. By segmenting the entire SAR image into superpixels and selecting long-term stable features such as urban areas, roads, grasslands, and bare soil for radiometric consistency evaluation. The evaluation results show that the average energy standard deviation for urban areas is 0.9524, and for bare soil is 0.4821, indicating higher radiometric consistency for single stars under the same incidence angle. When verifying the radiometric consistency among multiple stars using Sentinel-1 and TerraSAR satellites, this paper employs the Ulaby model to correct the backscattering coefficient of the TerraSAR satellite, reducing the calibration error from 1.78 dB to 0.63 dB. Through this correction, the impact of observation angle differences on the backscattering coefficient of SAR data is eliminated, ensuring that the radiometric consistency evaluation results are only influenced by the accuracy of satellite radiometric calibration, thus making the evaluation results more reliable.
|
