The current limited spaceborne hardware resources and the diversity of ship target scales in SAR images have led to the requirement of on-orbit real-time detection of ship targets in spaceborne synthetic aperture radar (SAR) images. In this paper, we propose a lightweight ship detection network based on the YOLOv4-LITE model. In order to facilitate the network migration to the satellite, the method uses MobileNetv2 as the backbone feature extraction network of the model. To solve the problem of ship target scale diversity in SAR images, an improved receptive field block (RFB) structure is introduced, enhancing the feature extraction ability of the network, and improving the accuracy of multi-scale ship target detection. A sliding window block method is designed to detect the whole SAR image, which can solve the problem of image input. Experiments on the SAR ship dataset SSDD show that the detection speed of the improved lightweight network could reach up to 47.16 FPS, with the mean average precision (mAP) of 95.03%, and the model size is only 49.34 M, which demonstrates that the proposed network can accurately and quickly detect ship targets. The proposed network model can provide a reference for constructing a spaceborne real-time lightweight ship detection network, which can balance the detection accuracy and speed of the network.

Himawari-8 (H8), as a geostationary satellite, observing the full disk image of the earth per 10 min, is a very powerful tool to investigate atmospheric aerosol temporal variations. The advanced himawari imager (AHI) onboard H8 has several spectral bands similar with those of Moderate Resolution Imaging Spectroradiometer (MODIS). Aerosol optical depth (AOD) is thus possible to be retrieved from AHI data using the dark target (DT) method of MODIS. Based on a statistics study of infrared bands of these two satellite instruments, we found that the difference of shortwave infrared bands between AHI (2.3 μm) and MODIS (2.12 μm) in DT area can be ignored for our purpose. Meanwhile, for AHI sensor, the normalized difference vegetation index (NDVI) calculated from 0.86 μm (instead of traditional 1.24 μm) and 2.3 μm can also be a good indicator of surface spectral reflectance. Therefore, first, a new NDVI of AHI sensor is proposed in this paper to improve the estimation of surface reflectance based on DT retrieval method. Then, AOD over DT area was retrieved using a lookup table strategy and tested based on AHI observation over China region from August 2015 to July 2016. All retrieved results are validated against ground-based measurements of the aerosol robotic network and the sun-sky radiometer observation network. The coincident AOD results between AHI retrieval- and ground-based network measurements show good agreements with R² of 0.81. The intercomparison between AHI and MODIS DT AOD products also shows a good agreement with R² of 0.92.

Columnar mass concentrations of aerosol components over the Arctic are estimated using microphysical parameters derived from direct sun extinction and sky radiance measurements of Aerosol Robotic Network. Aerosol optical, microphysical, chemical and radiative properties show that Arctic aerosols are dominated by fine mode particles, especially for high aerosol load cases. The average aerosol optical depth (AOD) of the selected Arctic sites in the sampling period is approximately 0.08, with 75% composed of fine mode particles. The fine mode fraction mostly exceeds 0.9 when AOD greater than 0.4. The ammonium sulfate-like component (AS) contributes about 68% of total dry aerosol mass for high-AOD events. The estimated compositions and back trajectories show that the transported aerosol particles from biomass burning events have large amounts of black carbon (BC) and brown carbon, while those from pollution events are characterised by large AS fractions. The instantaneous radiative forcing at the top-of-atmosphere is higher for the more absorbing components, and varies greatly with surface albedo and solar zenith angle. A regression model of columnar composition and radiative forcing within the atmosphere (RF_ATM) for Arctic aerosol is established, showing that BC dominates a positive RF_ATM with a high warming efficiency.