Search for: All records

Synthetic aperture radar (SAR) image classification is a challenging problem due to the complex imaging mechanism as well as the random speckle noise, which affects radar image interpretation. Recently, convolutional neural networks (CNNs) have been shown to outperform previous state-of-the-art techniques in computer vision tasks owing to their ability to learn relevant features from the data. However, CNNs in particular and neural networks, in general, lack uncertainty quantification and can be easily deceived by adversarial attacks. This paper proposes Bayes-SAR Net, a Bayesian CNN that can perform robust SAR image classification while quantifying the uncertainty or confidence of the network in its decision. Bayes-SAR Net propagates the first two moments (mean and covariance) of the approximate posterior distribution of the network parameters given the data and obtains a predictive mean and covariance of the classification output. Experiments, using the benchmark datasets Flevoland and Oberpfaffenhofen, show superior performance and robustness to Gaussian noise and adversarial attacks, as compared to the SAR-Net homologue. Bayes-SAR Net achieves a test accuracy that is around 10% higher in the case of adversarial perturbation (levels > 0.05). 

Tropospheric ozone (O₃) pollution is known to damage vegetation, reducing photosynthesis and stomatal conductance, resulting in modified plant transpiration to the atmosphere. We use an Earth system model to show that global transpiration response to near‐present‐day surface tropospheric ozone results in large‐scale global perturbations to net outgoing long‐wave and incoming shortwave radiation. Our results suggest that the radiative effect is dominated by a reduction in shortwave cloud forcing in polluted regions, in response to ozone‐induced reduction in land‐atmosphere moisture flux and atmospheric humidity. We simulate a statistically significant response of annual surface air temperature of up to ~ +1.5 K due to this ozone effect in vegetated regions subjected to ozone pollution. This mechanism is expected to further increase the net warming resulting from historic and future increases in tropospheric ozone.