Recent studies have proven that synthetic aperture radar (SAR) automatic target recognition (ATR) models based on deep neural networks (DNN) are vulnerable to adversarial examples. However, existing attacks are easily failed in the case where adversarial perturbations cannot be fully fed to victim models. We call this situation perturbation offset. Moreover, since background clutter takes up most of the areas in SAR images and has low relevance to recognition results, fooling models with global perturbations is quite inefficient. This paper proposes a semi-whitebox attack network, called Universal Local Adversarial Network (ULAN), to generate universal adversarial perturbations (UAP) for the target regions of SAR images. In the proposed network, we calculate the model’s attention heatmaps through layer-wise relevance propagation (LRP), which is used to locate the target regions of SAR images that have high relevance to recognition results. In particular, we utilize a generator based on the U-Net to learn the mapping from noise to UAPs and craft adversarial examples by adding the generated local perturbations to target regions. Experiments indicate that the proposed method fundamentally prevents perturbation offset and achieves comparable attack performance to conventional global UAPs by perturbing only a quarter or less of SAR image areas.

The paper presents the main experimental capabilities of the automatic system for the study of the interaction between porous media and atmosphere. This system is one of the possible variants of patent 2-241-477. The paper shows the result of a drying and wetting study of a tailings sample. The sample is 32 cm in height and 29 cm in diameter. The saturation and drying process was carried out at 26 degrees Celsius at the sample surface, an air velocity of 2 m/s and a relative humidity of 60%. The recharge applied to the column surface was 0.3 cm3/cm2/minute. The results show the effectiveness of the system in controlling the spatial-temporal variation of different physical properties of the porous medium in the vertical and surface profile. The controlled variables include volumetric water content, suction, relative humidity, sample weight variation due to recharge or drying, shrinkage, temperature, etc. The experiment shows how the tailings saturate very quickly but dry very slowly. Tailings consolidation occurs in the drying and wetting processes, being higher in the first one.