Aircraft detection in synthetic aperture radar (SAR) images is still a challenging research task because of the insufficient public data, the difficulty of multiscale target detection, and the complexity of background interference. In this article, we construct a public SAR aircraft detection dataset (SADD) with complex background and interference objects to facilitate the research in SAR aircraft detection. Then, we propose the scale expansion and feature enhancement pyramid network as the SADD baseline. It uses a four-scale fusion method to combine the shallow position information with the deep semantic information, effectively adapting to the multiscale target detection in SAR images, significantly improving the detection effect of small targets. The feature enhancement pyramid structure is connected behind the backbone network to weaken the background texture and highlight the target to achieve feature enhancement, improving the ability to extract target features in complex backgrounds. Finally, to further improve the detection performance of the small-scale SAR aircraft dataset, we propose a domain adaptive transfer learning method. Experiments on SADD show that this method can significantly improve the recall rate and F1 score. At the same time, we find that the transfer effect of using homologous but different types of targets as the source domain is better than those of heterologous but same types of targets in SAR aircraft detection, which is instructive for future research.

Military-civilian attribute recognition of ships in synthetic aperture radar (SAR) imagery plays an important role in marine surveillance. However, high-quality labeled data are hard to obtain for SAR ships, which hinder the development of deep learning models. Considering that models directly transferred from labeled optical images cannot achieve satisfactory performance for SAR applications due to the great discrepancy of different modalities, we propose a two-stage transfer learning method by combining the data-level and feature-level knowledge transfer. First, CycleGAN is adopted in the first stage to transfer the labeled optical image domain to the intermediate SAR-like image domain with the attribute labels. Then, a novel network called Domain Transfer using Adversarial learning and Metric learning (DTAM) is proposed to realize the task of military-civilian ship recognition by the domain adaption of the intermediate domain and the target SAR domain with joint adversarial learning and metric learning. To validate the proposed method, we establish a high-resolution SAR ship recognition dataset (HRSSRD), containing SAR and optical images of military and civilian ships. The experimental results show that the proposed two-stage architecture exhibits promising performance on the problem of SAR military-civilian ship recognition.

Remote sensing image scene classification plays an important role in remote sensing image interpretation. Deep learning brings prosperity to the research in this field, and numerous deep learning models are proposed in order to improve the performance of scene classification. However, images of different remote sensing scenes vary a lot, showing similar or diverse textures and simple or complex contents. Using a fixed convolution neural network framework to classify scene images is performance-limited and not practice-flexible. To address this issue, in this paper, we propose the SEMSDNet (multi-scale dense networks with squeeze and excitation attention). The framework MSDNet with multiple classifiers and dense connections can automatically transform between a small network and a deep network according to the complexity of test samples and the limitation of computational resources. Moreover, in order to extract more effective features, the squeeze and excitation attention mechanism is introduced into the framework to process the features of various scenes self-adaptively. In addition, considering the limited computing resources, we impose two settings with computational constraints at the test time: budgeted batch classification which is a fixed computational budget setting for sample classification, and anytime prediction which forces the network to output a prediction at any given point-in-time. Experimental results on several public datasets show that the proposed SEMSDNet method is superior to state-of-the-art methods on both performance and efficiency. Experiments also reveal its capability to treat samples of different classification difficulties with uneven resource allocation and flexible network architecture, showing its potentials in practical applications.

Semantic segmentation is a fundamental task in remote sensing image interpretation, which aims to assign a semantic label for every pixel in the given image. Accurate semantic segmentation is still challenging due to the complex distributions of various ground objects. With the development of deep learning, a series of segmentation networks represented by fully convolutional network (FCN) has made remarkable progress on this problem, but the segmentation accuracy is still far from expectations. This paper focuses on the importance of class-specific features of different land cover objects, and presents a novel end-to-end class-wise processing framework for segmentation. The proposed class-wise FCN (C-FCN) is shaped in the form of an encoder-decoder structure with skip-connections, in which the encoder is shared to produce general features for all categories and the decoder is class-wise to process class-specific features. To be detailed, class-wise transition (CT), class-wise up-sampling (CU), class-wise supervision (CS), and class-wise classification (CC) modules are designed to achieve the class-wise transfer, recover the resolution of class-wise feature maps, bridge the encoder and modified decoder, and implement class-wise classifications, respectively. Class-wise and group convolutions are adopted in the architecture with regard to the control of parameter numbers. The method is tested on the public ISPRS 2D semantic labeling benchmark datasets. Experimental results show that the proposed C-FCN significantly improves the segmentation performances compared with many state-of-the-art FCN-based networks, revealing its potentials on accurate segmentation of complex remote sensing images.

Detecting urban areas from very high resolution (VHR) remote sensing images plays an important role in the field of Earth observation. The recently-developed deep convolutional neural networks (DCNNs), which can extract rich features from training data automatically, have achieved outstanding performance on many image classification databases. Motivated by this fact, we propose a new urban area detection method based on DCNNs in this paper. The proposed method mainly includes three steps: (i) a visual dictionary is obtained based on the deep features extracted by pre-trained DCNNs; (ii) urban words are learned from labeled images; (iii) the urban regions are detected in a new image based on the nearest dictionary word criterion. The qualitative and quantitative experiments on different datasets demonstrate that the proposed method can obtain a remarkable overall accuracy (OA) and kappa coefficient. Moreover, it can also strike a good balance between the true positive rate (TPR) and false positive rate (FPR).

Vehicle detection is a very important application of remote sensing. However, suffering from the low acutance and insufficient color information, the detection of weak vehicles in satellite imagery still remains a challenge. Image enhancement can improve the visual effects of remote sensing images. Nevertheless, most existing image enhancement methods aim to improve the quality of the entire image without target guidance, which have ambiguous contributions to the detection performance. Methods based on generative adversarial networks (GANs) have realized image enhancement with target guidance by the addition of target-guided branches, but paired training data is not available in some scenarios. In this article, a novel model of detection-guided CycleGAN (DE-CycleGAN) is proposed to enhance the weak targets for the purpose of accurate vehicle detection, where a backbone GAN with a target-guided branch is learned in the absence of paired images. Specifically, enhancements of two levels are mutually executed. At the image level, the color information of the entire satellite image is enriched by refined CycleGAN, and its sharpness is enhanced by the gradient enhancement model. At the object level, the target-guided branch for detection is added to enhance features of the target. The experimental results validate that the detection performance has been significantly improved on the images enhanced by the proposed DE-CycleGAN model, which shows a positive effect on weak target detection.