L band aperture synthesis microwave Radiometer (LASMR) on board Chinese Ocean Salinity Mission(COSM) is a two-dimensional interferometer, which, due to the Fourier-transform relationship between the spatial domain and the brightness temperature, possesses a well-behaved redundancy characteristic. The analysis results show that when the failed units are located on long baselines, the more units that fail, the greater the decrease in system resolution, while if the failed units are on short baselines, the impact on system resolution is relatively minor. LASMR improves resolution by synthesizing the aperture but worsens sensitivity, hence when there are unit failures, the system’s noise floor is reduced in most cases, showing a trend towards optimization. The absence of long and medium baselines has a minor impact on the quality of the brightness temperature image, while the lack of short baselines significantly affects the image quality, leading to a rapid deterioration in image accuracy. However, under fixed unit failure conditions, the impact on images for different targets is biased in effect. The error correction of the brightness temperature can be performed using the ocean target transform (OTT) technique, and the image accuracy is significantly improved. Even under the adverse condition of 15 unit failures, the image accuracy can still reach 0.12K. The redundancy characteristics of the LASMR are evident.
In order to improve the quality and efficiency of the bright temperature reconstruction image of the synthetic aperture microwave radiometer, this paper takes the gray value of remote sensing image with channel amplitude and phase error and random error as the original bright temperature image, uses convolutional neural network (CNN) to supervise and learn the mapping relationship between bright temperature image and visibility function, and reconstructs bright temperature image according to the learned mapping relationship. It is compared with the traditional hexagon Fourier transform(HFFT). In terms of visual effect, the CNN network inversion method has clearer boundary and better effect. In terms of evaluation indexes, RMSE values of HFFT inversion method and CNN network inversion method are 15.80K and 10.93K, and PSNR values are 23.88dB and 27.09dB, respectively. Therefore, compared with HFFT inversion method, CNN network inversion method has less error in reconstruction results, effectively reduces Gibbs effect, and can better restore the original bright temperature image.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.