The object detection technologies of remote sensing are widely used in various fields, such as environmental monitoring, geological disaster investigation, urban planning, and military defense. However, the detection algorithms lack the robustness to detect tiny objects against complex backgrounds. In this paper, we propose a Multiple Attention Mechanism Enhanced YOLOX (MAME-YOLOX) algorithm to address the above problem. Firstly, the CBAM attention mechanism is introduced into the backbone of the YOLOX, so that the detection network can focus on the saliency information. Secondly, to identify the high-level semantic information and enhance the perception of local geometric feature information, the Swin Transformer is integrated into the YOLOX’s neck module. Finally, instead of GIOU loss, CIoU loss is adopted to measure the bounding box regression loss, which can prevent the GIoU from degenerating into IoU. The experimental results of three publicly available remote sensing datasets, namely, AIBD, HRRSD, and DIOR, show that the algorithm proposed possesses better performance, both in relation to quantitative and qualitative aspects.

The current popular one-shot multi-object tracking (MOT) algorithms are dominated by the joint detection and embedding paradigm, which have high inference speeds and accuracy, but their tracking performance is unstable in crowded scenes. Not only does the detection branch have difficulty in obtaining the accurate object position, but the ambiguous appearance of features extracted by the re-identification (re-ID) branch also leads to identity switches. Focusing on the above problems, this paper proposes a more robust MOT algorithm, named CSMOT, based on FairMOT. First, on the basis of the encoder–decoder network, a coordinate attention module is designed to enhance the information interaction between channels (horizontal and vertical coordinates), which improves its object-detection abilities. Then, an angle-center loss that effectively maximizes intra-class similarity is proposed to optimize the re-ID branch, and the extracted re-ID features are made more discriminative. We further redesign the re-ID feature dimension to balance the detection and re-ID tasks. Finally, a simple and effective data association mechanism is introduced, which associates each detection instead of just the high-score detections during the tracking process. The experimental results show that our one-shot MOT algorithm achieves excellent tracking performance on multiple public datasets and can be effectively applied to crowded scenes. In particular, CSMOT decreases the number of ID switches by 11.8% and 33.8% on the MOT16 and MOT17 test datasets, respectively, compared to the baseline.

To detect faint target stars of 22nd magnitude and above, an astronomical exploration project requires its space camera’s readout noise to be less than 5e⁻ with long-time working stability. Due to the limitation of satellite, the traditional CCD detector-based camera does not meet the requirements, including volume, weight, and power consumption. Thereby, a low-noise ultra-stable camera based on 9 K × 9 K large target surface CMOS is designed to meet the needs. For the first time, the low-noise ultra-stable camera based on CMOS detector will be applied to space astronomy projects, remote sensing imaging, resource survey, atmospheric and oceanic observation and other fields. In this paper, the design of the camera is introduced in detail, and the camera is tested for several rounds at −40 °C; it also undergoes further testing and data analysis. Tests proved super stability and that the readout noise is lower than 4.5e⁻. Dark current, nonlinearity and PTC indicators meet the requirements of the astronomical exploration project.

In order to solve the star identification problem in the lost space mode for scientific cameras with small fields of view and higher instruction magnitudes, this paper proposes a star identification algorithm based on a hybrid grid pattern. The application of a hybrid pattern generated by multi-calibration stars in the initial matching enables the position distribution features of neighboring stars around the main star to be more comprehensively described and avoids the interference of position noise and magnitude noise as much as possible. Moreover, calibration star filtering is adopted to eliminate incorrect candidates and pick the true matched navigation star from candidate stars in the initial match. Then, the reference star image is utilized to efficiently verify and determine the final identification results of the algorithm via the nearest principle. The performance of the proposed algorithm in simulation experiments shows that, when the position noise is 2 pixels, the identification rate of the algorithm is 96.43%, which is higher than that of the optimized grid algorithm by 2.21% and the grid algorithm by 4.05%; when the magnitude noise is 0.3 mag, the star identification rate of the algorithm is 96.45%, which is superior to the optimized grid algorithm by 2.03% and to the grid algorithm by 3.82%. In addition, in the actual star image test, star magnitude values of ≤12 mag can be successfully identified using the proposed algorithm.