Image super-resolution (SR) technique can improve the spatial resolution of images without upgrading the imaging system. As a result, SR promotes the development of high resolution (HR) remote sensing image applications. Many remote sensing image SR algorithms based on deep learning have been proposed recently, which can effectively improve the spatial resolution under the constraints of HR images. However, images acquired by remote sensing imaging devices typically have lower resolution. Hence, an insufficient number of HR remote sensing images are available for training deep neural networks. In view of this problem, we propose an unsupervised SR method that does not require HR remote sensing images. The proposed method introduces a generative adversarial network (GAN) that obtains SR images through the generator; then, the SR images are downsampled to train the discriminator with low resolution (LR) images. Our method outperformed several methods in terms of the quality of the obtained SR images as measured by 6 evaluation metrics, which proves the satisfactory performance of the proposed unsupervised method for improving the spatial resolution of remote sensing images.

Infrared small target detection is widely used in the military field, and robust infrared small target detection has significant significance. Inspired by plants, an infrared small target detection method based on the four-leaf model is proposed. This model has both macro and micro attributes, with macro attributes referred to as the background suppressor (BS) and micro attributes referred to as the texture collector (TC). BS is a four-neighborhood model that can achieve background suppression while reducing the interference of bright background clutter in the target neighborhood to a certain extent. TC can collect texture information of small targets and improve the enhancement effect of small targets. The fusion of TC and BS can effectively suppress background clutter and improve the detection performance of infrared small targets. The experiment is carried out on five real infrared image sequences. The results show that the proposed infrared small target detection method can improve the detection rate and reduce the false alarm rate in the face of infrared images with complex backgrounds. Compared to existing algorithms, the algorithm has high robustness.

In remote sensing, it is desirable to improve image resolution by using the image super-resolution (SR) technique. However, there are two challenges: the first one is that high-resolution (HR) images are insufficient or unavailable; another one is that the single degradation model such as bicubic (BIC) cannot super-resolve favorable images in the real world. To address the above two problems, this article presents a multi-degradation, unsupervised SR method based on deep learning. This framework consists of a degrader D to fit the image degradation model and a generator G to generate SR image. By introducing D, calculating the loss function between SR image and HR image as supervised SR methods did can be converted into calculating loss between low resolution (LR) image and image degraded by SR image, thereby realizing unsupervised learning. Experiments on several degradation models show that our method renders the state-of-the-art results compared with existing unsupervised SR methods, and achieves competitive results in contrast with supervised SR methods. Moreover, for real remote sensing images obtained by the Jilin-1 satellite, our method obtained more plausible results visually, which demonstrate the potential in real-world applications.

Realizing high robust infrared small target detection is of great significance in the face of complex and ever-changing infrared image sequences. In this letter, we analyze the inherent relationship between local contrast, local gradient, and local background gray value of infrared small targets from a new perspective. Based on this, a method for infrared small target detection based on local contrast and the local gradient is proposed. In addition, a four-direction model is proposed to reduce the interference of bright background clutter in the target neighborhood. The experimental results show that the method proposed in this paper has high robustness and is suitable for detecting small infrared targets in complex backgrounds.

Marine oil spills accidents has caused a large amount of crude oil to leak into the marine environment and threaten marine ecological environment. Hyperspectral remote sensing images(HRSI) technology can quickly and accurately identify oil film of different thickness on marine surface. In order to overcome the traditional spectrum analysis method and space extraction method of long time sampling, calculation, analysis and other shortcomings. On account of the advantages of the spectral and spatial information in the field of HRSI classification, a spectral-spatial features extraction(SSFE) method based convolutional neural networks(CNNs) was proposed to analyse oil spills. In this way, one and two dimensional models based on convolutional neural networks(1D-CNN,2D-CNN) have been introduced as the spectral and spatial features extractor. When extracting spatial features, double-two convolution layers are connected to increasing nonlinearity and reduce the number of parameters. Furthermore, in order to address overfitting and imbalance samples, L2 regularization, class_weight and drouput is added to classes data modeling. More importantly, principal component analysis (PCA) is applied to data dimension reduction, 1D-CNN and 2D-CNN is combined into a unified model further extract the joint spatial-spectral features. To evaluate the effectiveness of the proposed approach, three hyperspectral datasets were utilized, including: University of Pavia dataset, oil spill area 1, oil spill area 2. Experimental results reveal that the proposed method have a very satisfactory performance and better distinguish oil spills.

To improve the phase-shifting accuracy, this paper presents a novel integrated framework for design, control and experimental validation of the piezoelectric actuated phase shifter with a trade-off between accuracy and cost. The piezoelectric actuators with built-in sensors are adopted to drive the double parallel four-bar linkage flexure hinge-based mechanisms. Three mechanisms form the tripod structure of the assembled phase shifter. Then, a semi-closed loop controller with inner feedback and outer feedforward loops via the external laser interferometer is developed for accurate positioning of the phase shifter. Finally, experiments related with travel range, step response, linearity and repeatability are carried out. The linearity error is 0.21% and the repeatability error of 10 $\mathsf{\mu }$ m displacement is 3 nm. The results clearly demonstrate the good performance of the developed phase shifter and the feasibility of the proposed integrated framework.