Ship detection of synthetic aperture radar (SAR) images is one of the research hotspots in the field of marine surveillance. Fusing salient features to detection network can effectively improve the precision of ship detection. However, how to effectively fuse the salient features of SAR images is still a difficult task. In this paper, to improve the ship detection precision, we design a novel one-stage ship detection network to fuse salient features and deep convolutional neural network (CNN) features. Firstly, a saliency map extraction algorithm is proposed. The algorithm is applied to generate saliency map by using multi-scale pyramid features and frequency domain features. Secondly, the backbone of the ship detection network contains a two-stream network. The upper-stream network uses the original SAR image as input to extract multi-scale deep CNN features. The lower-stream network uses the corresponding saliency map as input to acquire multi-scale salient features. Thirdly, for integrating the salient features to deep CNN features, a novel salient feature fusion method is designed. Finally, an improved bi-directional feature pyramid network is applied to the ship detection network for reducing the computational complexity and network parameters. The proposed methods are evaluated on the public ship detection dataset and the experimental results shows that it can make a significant improvement in the precision of SAR image ship detection.

Synthetic aperture radar (SAR)¹ is an active Earth observation system deployed on aircraft, satellites, or other flight platforms. Compared with the optical and infrared systems, SAR can provide all-time, all-weather, high-resolution, and wide-swath observation. It also has a certain ability to penetrate the Earth’s surface and discover underground targets. Therefore, the SAR has advantages in disaster monitoring,² environmental monitoring,³ ocean surveillance,⁴ resource exploration,⁵ surveying, and military applications.

This paper focuses on the trajectory planning that optical satellites approach the geostationary (GEO) spacecraft and the trajectory planning that the optical satellite flies around the GEO spacecraft for continuous surveillance of the GEO spacecraft The optical satellite approaches the GEO spacecraft with continuous thrust and flies around the GEO spacecraft without thrust. The observation angles of the optical satellite to the GEO spacecraft are less than 30 degrees throughout the entire fly-around mission. Two methods are used in fly-around trajectory planning: trajectory planning based on classical orbital elements and trajectory planning based on CW equations. The advantage of trajectory planning based on classical orbital elements is that the position of the optical satellite relative to the GEO spacecraft is clear in earthcentered-inertial frame and the calculation method of fly-around orbital elements is simple. If the orbital inclination of the GEO spacecraft is less than 0.02 degrees and the semi-major axis of the elliptical relative trajectory of the fly-around formation is greater than 60 km, fly-around trajectory planning based on classical orbital elements is more efficient and easier than that based on CW equations. Optimal control theory is used in the trajectory planning for optical satellites approaching GEO spacecraft and the trajectory optimization is a two-point boundary-value problem. Simulations have proved that this method is feasible, which is highly effective in engineering. This method can be used for space situational awareness and on-orbit servicing missions of GEO spacecraft.

The solution of the short-arc angles-only orbit determination problem has large uncertainty because the topocentric range is not observable. For a certain angular observation tracklet with measurement noise, there exist numerous potential orbits, all of which are compatible with the observations. However, the solution of a deterministic initial orbit determination algorithm is usually far different from the true orbit, especially for the semimajor axis and the eccentricity. A new sampling method is proposed to describe the probability distribution of the orbit determination solutions. Firstly, a series of orbits are sampled in the semimajor axis - eccentricity plane. A chi-square test method is proposed to select candidate orbits from the sample orbits. The weights of the candidate orbits are calculated to measure their probability being the true orbit. Finally, the kernel density estimation algorithm is used to estimate the probability density function of the true orbits. With some a priori assumptions, the candidate orbits can be further screened, and their weight can be modified. The a priori knowledge can significantly improve the accuracy of the orbit determination solution.

Self-supervised method has proven to be a suitable approach for despeckling on synthetic aperture radar (SAR) images. However, most self-supervised despeckling methods are trained by noisy-noisy image pairs, which are constructed by using natural images with simulated speckle noise, time-series real-world SAR images or generative adversarial network, limiting the practicability of these methods in real-world SAR images. Therefore, in this paper, a novel self-supervised despeckling algorithm with an enhanced U-Net is proposed for real-world SAR images. Firstly, unlike previous self-supervised despeckling works, the noisy-noisy image pairs are generated from real-word SAR images through a novel generation training pairs module, which makes it possible to train deep convolutional neural networks using real-world SAR images. Secondly, an enhanced U-Net is designed to improve the feature extraction and fusion capabilities of the network. Thirdly, a self-supervised training loss function with a regularization loss is proposed to address the difference of target pixel values between neighbors on the original SAR images. Finally, visual and quantitative experiments on simulated and real-world SAR images show that the proposed algorithm notably removes speckle noise with better preserving features, which exceed several state-of-the-art despeckling methods.

This paper proposes an improved sampling-based approach for spacecraft proximity operation path planning under Clohessy-Wiltshire-Hill dynamics. The proposed approach is based on a modified version of the FMT* (Fast Marching Tree) algorithm with safety strategy which is divided into three parts: (1) incorporating relative ellipse to simplify the sampling state space and avoid collision with target; (2) combining internal/external ellipse-based collision detection algorithms for hovering obstacle and non-coplanar ellipse obstacle; (3) using rotating hyperplane method to handle the coplanar ellipse obstacle and uncertainty obstacle. By referring the safety strategy to simplify the state space before FMT* algorithm, the approach can reduce the complexity of path planning, especially the resampling and collision avoidance detection cyclic process in FMT*, thereby improve the planning efficiency. Two simulated scenarios, a coplanar path planning with/without coplanar ellipse obstacle, are developed to illustrate the approach. As a result, the proposed approach appears to be potential for spacecraft proximity real-time path planning as well as other complex space mission path planning generalized to different dynamics and environments, such as On-Orbit Service, and enabling a real-time computation of low-cost trajectory.