Deep learning-based synthetic aperture radar (SAR) target recognition often suffers from sparsely distributed training samples and rapid angular variations due to scattering scintillation. Thus, data-driven SAR target recognition is considered a typical few-shot learning (FSL) task. This article first reviews the key issues of FSL and provides a definition of the FSL task. A novel adversarial autoencoder (AAE) is then proposed as an SAR representation and generation network. It consists of a generator network that decodes target knowledge to SAR images and an adversarial discriminator network that not only learns to discriminate “fake” generated images from real ones but also encodes the input SAR image back to target knowledge. The discriminator employs progressively expanding convolution layers and a corresponding layer-by-layer training strategy. It uses two cyclic loss functions to enforce consistency between the inputs and outputs. Moreover, rotated cropping is introduced as a mechanism to address the challenge of representing the target orientation. The moving and stationary Target recognition (MSTAR) 7-target dataset is used to evaluate the AAE’s performance, and the results demonstrate its ability to generate SAR images with aspect angular diversity. Using only 90 training samples with at least 25° of orientation interval, the trained AAE is able to generate the remaining 1748 samples of other orientation angles with an unprecedented level of fidelity. Thus, it can be used for data augmentation in SAR target recognition FSL tasks. Our experimental results show that the AAE could boost the test accuracy by 5.77%.

A zero-shot learning (ZSL) method of automatic target recognition (ATR) in synthetic aperture radar (SAR) image is proposed to address the scenario, where no SAR sample of a particular target is available for training. To learn features of the unseen target, physics-based electromagnetic (EM) simulated images of the target under different azimuth angles are used as the training data instead. The challenge lies in the fact that the simulated image has a distinct but nonessential texture that the real images do not have and, thus, can easily result in an overfitted discriminator network. To overcome this problem, all images are first preprocessed with a nonessential factor suppression step and then fed into a pretrained convolutional neural network for feature extraction. Finally, the feature vector is fed into a trainable fully-connected network for classification. The low-dimensional embedding of feature vectors suggests that the nonessential factor suppression can align the simulated samples with true samples effectively. We propose the max-tolerability principle and averaged margin index for ZSL, which is a useful indicator for selecting optimal classifier. We validated our method on ten-type target recognition task on MSTAR data sets and achieved 91.93% accuracy on nine known targets and 79.08% accuracy on zero-shot target.

Earth observation data have huge potential to enrich our knowledge about our planet. An important step in many Earth observation tasks is semantic segmentation. Generally, a large number of pixelwise labeled images are required to train deep models for supervised semantic segmentation. On the contrary, strong intersensor and geographic variations impede the availability of annotated training data in Earth observation. In practice, most Earth observation tasks use only the target scene without assuming availability of any additional scene, labeled or unlabeled. Keeping in mind such constraints, we propose a semantic segmentation method that learns to segment from a single scene, without using any annotation. Earth observation scenes are generally larger than those encountered in typical computer vision datasets. Exploiting this, the proposed method samples smaller unlabeled patches from the scene. For each patch, an alternate view is generated by simple transformations, e.g., addition of noise. Both views are then processed through a two-stream network and weights are iteratively refined using deep clustering, spatial consistency, and contrastive learning in the pixel space. The proposed model automatically segregates the major classes present in the scene and produces the segmentation map. Extensive experiments on four Earth observation datasets collected by different sensors show the effectiveness of the proposed method. Implementation is available at https://gitlab.lrz.de/ai4eo/cd/-/tree/main/unsupContrastiveSemanticSeg .

Subtle volcanic deformations point to volcanic activities, and monitoring them helps predict eruptions. Today, it is possible to remotely detect volcanic deformation in mm/year scale thanks to advances in interferometric synthetic aperture radar (InSAR). This article proposes a framework based on a deep learning model to automatically discriminate subtle volcanic deformations from other deformation types in five-year-long InSAR stacks. Models are trained on a synthetic training set. To better understand and improve the models, explainable artificial intelligence (AI) analyses are performed. In initial models, Gradient-weighted Class Activation Mapping (Grad-CAM) linked new-found patterns of slope processes and salt lake deformations to false-positive detections. The models are then improved by fine-tuning (FT) with a hybrid synthetic-real data, and additional performance is extracted by low-pass spatial filtering (LSF) of the real test set. The t-distributed stochastic neighbor embedding (t-SNE) latent feature visualization confirmed the similarity and shortcomings of the FT set, highlighting the problem of elevation components in residual tropospheric noise. After fine-tuning, all the volcanic deformations are detected, including the smallest one, Lazufre, deforming 5 mm/year. The first time confirmed deformation of Cerro El Condor is observed, deforming 9.9–17.5 mm/year. Finally, sensitivity analysis uncovered the model’s minimal detectable deformation of 2 mm/year.

Deep transfer-learning-based change detection methods are dependent on the availability of sensor-specific pretrained feature extractors. Such feature extractors are not always available due to lack of training data, especially for hyperspectral sensors and other hyperdimensional images. Moreover models trained on easily available multispectral (RGB/RGB-NIR) images cannot be reused on such hyperdimensional images due to their irregular number of bands. While hyperdimensional images show large number of spectral bands, they generally show much less spatial complexity, thus reducing the requirement of large receptive fields of convolution filters. Recent works in the computer vision have shown that even untrained deep models can yield remarkable result in some tasks like super-resolution and surface reconstruction. This motivates us to make a bold proposition that untrained lightweight deep model, initialized with some weight initialization strategy, can be used to extract useful semantic features from bi-temporal hyperdimensional images. Based on this proposition, we design a novel change detection framework for hyperdimensional images by extracting bitemporal features using an untrained model and further comparing the extracted features using deep change vector analysis to distinguish changed pixels from the unchanged ones. We further use the deep change hypervectors to cluster the changed pixels into different semantic groups. We conduct experiments on four change detection datasets: three hyperspectral datasets and a hyperdimensional polarimetric synthetic aperture radar dataset. The results clearly demonstrate that the proposed method is suitable for change detection in hyperdimensional remote sensing data.

Measuring and mapping vegetation structure is essential for understanding the functioning of terrestrial ecosystems and for informing environmental policies. Recent years have seen a growing demand for high‐resolution data on vegetation structure, driving their prediction at fine resolutions (1–30 m) at state, continental, and global spatial extents by combining satellite data with machine learning. As these initiatives expand, it is crucial to actively discuss the quality and usability of these products. Here, we briefly summarize current efforts to map vegetation structure and show that continental‐to‐global canopy height models (CHMs) exhibit significant errors in canopy heights compared to national airborne laser scanning (ALS) data. We recommend that regions with abundant ALS data, such as Europe, prioritize using ALS‐based canopy height metrics rather than relying on less accurate predictions from satellite products. Despite variations in ALS data characteristics, such as temporal inconsistencies and differences in acquisition characteristics and classification accuracy, the generation of spatially contiguous canopy height products in raster format at fine spatial resolution is necessary and feasible. This requires coordinating efforts for data and survey harmonization, developing standardized processing pipelines and continent‐wide ALS products, and ensuring free access for research and environmental policy. We show that ALS data now cover most of Europe, with newer surveys achieving higher point densities, improving their suitability for vegetation mapping. Beyond numerous applications in forestry, ecology, and conservation, such data sets are crucial for calibrating future Earth Observation missions, making them essential for producing reliable and accurate global, fine‐resolution vegetation structure data.