Multi-label classification in remote sensing is very significant which plays an important role in extracting valuable information from satellite imagery. Ignoring the distinct information provided by labels in each image or transforming images into content-aware category representations without considering inherent correlation of labels within the dataset can result in the establishment of improper relationships between images and labels, ultimately leading to a significant degradation in accuracy. To address this problem, this paper proposes a Label-Driven Graph Convolutional Network (LD-GCN) to excavate substantial information using the inherent correlation of labels from datasets and build a strong relationship between labels and images. The framework consists of two modules, i.e. label recognition GCN (LRGCN) and semantic enrichment module (SEM). The LRGCN module yields rich and valuable information from the inherent correlation of labels and builds a strong relationship between images and labels. The SEM further enriches the semantics obtained from LRGCN. Experiments conducted on UCM, AID, and DFC15 multi-label remote sensing datasets illustrate that LD-GCN outperforms the state-of-the-art methods on key evaluation metrics.

Imaging Radar Altimeter (IRA) is the current development tendency for ocean surface topography (OST) altimetry, which utilizes Synthetic Aperture Radar (SAR) and interferometry to improve the spatial resolution of OST to several kilometers or even better. Meanwhile, centimetric altimetry accuracy should be guaranteed for applications such as geostrophic currents or marine gravity anomaly inversion. However, the baseline length of IRA which determines the altimetric sensitivity is confined by the satellite platform, in consideration of baseline vibration and payload capability. Therefore, the baseline length from a single satellite can extend to only tens of meters, making it difficult to achieve centimetric accuracy. Referring to the successful experience from TerraSAR-X/TanDEM-X, satellite formation can easily extend the baseline length to hundreds or thousands of meters, depending on the helix orbit. Therefore, we propose the large baseline IRA (LB-IRA) from satellite formation for OST altimetry: the carrier frequency shift (CFS) is brought in to compensate for the severe baseline decorrelation, and the helix orbit is carefully selected to prevent severe time decorrelation from along-track baseline. The numerical results indicate that the LB-IRA, whose cross-track baseline ranges between 629~1000 m and along-tack baseline ranges between 0~40 m, can achieve ~1 cm relative accuracy at 1 km resolution.

Low Earth orbit (LEO) satellite constellations have emerged as an effective alternative for the provision of high-accuracy positioning, navigation and timing (PNT) solutions which are based on high-precision orbit and clock information. Determining an orbit with high precision is dependent on the number and distribution of ground tracking stations. Therefore, it is important to investigate methodologies that can ensure the adequate observing coverage of LEO navigation constellations. In this study, an evolutionary algorithm is applied to optimize the number and deployment of ground stations for tracking LEO constellations. According to the distribution area, two schemes of study are analyzed: (a) global deployment—the ground stations are deployed throughout the globe; (b) regional deployment—a selected region is used for deployment. For global deployment, the optimization objectives are focused on the ground station and observing rate for k-heavy observing coverage (HC), while the sole objective for the regional deployment scheme is the satellite position dilution of precision (SPDOP). It is shown that a deployment of 95 ground stations is optimal for achieving 3-HC with an observing rate of 97.37% and 4-HC with an observing rate of 92.01%. For regional distribution, 15, 20 and 25 ground stations are used for three optimal configurations of SPDOP at 2.058, 1.399 and 1.330, respectively. The results are significantly enhanced using intersatellite links for SPDOP evaluation, from 2.058, 1.399 and 1.330 to 0.439, 0.422 and 0.409, with 15, 20 and 25 ground stations, respectively.

Object detection in remote sensing images plays an important role in both military and civilian remote sensing applications. Objects in remote sensing images are different from those in natural images. They have the characteristics of scale diversity, arbitrary directivity, and dense arrangement, which causes difficulties in object detection. For objects with a large aspect ratio and that are oblique and densely arranged, using an oriented bounding box can help to avoid deleting some correct detection bounding boxes by mistake. The classic rotational region convolutional neural network (R2CNN) has advantages for text detection. However, R2CNN has poor performance in the detection of slender objects with arbitrary directivity in remote sensing images, and its fault tolerance rate is low. In order to solve this problem, this paper proposes an improved R2CNN based on a double detection head structure and a three-point regression method, namely, TPR-R2CNN. The proposed network modifies the original R2CNN network structure by applying a double fully connected (2-fc) detection head and classification fusion. One detection head is for classification and horizontal bounding box regression, the other is for classification and oriented bounding box regression. The three-point regression method (TPR) is proposed for oriented bounding box regression, which determines the positions of the oriented bounding box by regressing the coordinates of the center point and the first two vertices. The proposed network was validated on the DOTA-v1.5 and HRSC2016 datasets, and it achieved a mean average precision (mAP) of 3.90% and 15.27%, respectively, from feature pyramid network (FPN) baselines with a ResNet-50 backbone.

A diffractive imaging system consisting of two satellites is analyzed in view of dynamics. The mathematical model of rigid and flexion couples is studied to describe the relative motion of diffractive satellites and imaging satellites. Based on an integrated dynamics model with dual quaternion, a fixed-time non-singular terminal sliding mode controller is designed to meet the requirements of Earth observation. Finally, introducing the non-singular terminal sliding mode as the control group, a comparative simulation of relative motion and control is implemented to verify the controller and dynamics model.