National University Of Defense Technology Research Articles

Infrared Weak and Small Target Detection Algorithm Based on Deep Learning

Abstract In the infrared imaging scene where the target is at a long distance and the background is cluttered, due to the interference of noise and background texture information, the infrared image is prone to problems such as low contrast between the target and the background, and feature confusion, which makes it difficult to accurately extract and detect the target. To solve this problem, firstly, the infrared image is enhanced by combining DDE and MSR algorithm to improve the contrast and detail visibility of the image. For the RT-DETR network structure, the EMA attention mechanism is introduced into the backbone to enhance the feature extraction ability of the model by extracting context information. The CAMixing convolutional attention module is introduced into CCFM, and the multi-scale convolutional self-attention mechanism is introduced to focus on local information and enhance the detection ability of small targets. The filtering rules of the prediction box are improved, combined with Shape-IoU, and the convergence speed of the loss function in the detection and the detection accuracy of small targets are improved by paying attention to the influence of the intrinsic properties of the bounding box itself on the regression. In the experiment, the infrared weak target image dataset of the National University of Defense Technology was selected, labeled and trained. Experimental results show that compared with the original DETR algorithm, the average precision of the improved algorithm (mAP) is increased by 3.2%, and it can effectively detect infrared weak and small targets in different complex backgrounds, which reflects good robustness and adaptability, and can be effectively applied to infrared weak and small target detection in complex backgrounds.

CenterADNet: Infrared Video Target Detection Based on Central Point Regression.

Infrared video target detection is a fundamental technology within infrared warning and tracking systems. In long-distance infrared remote sensing images, targets often manifest as circular spots or even single points. Due to the weak and similar characteristics of the target to the background noise, the intelligent detection of these targets is extremely complex. Existing deep learning-based methods are affected by the downsampling of image features by convolutional neural networks, causing the features of small targets to almost disappear. So, we propose a new infrared video weak-target detection network based on central point regression. We focus on suppressing the image background by fusing the different features between consecutive frames with the original image features to eliminate the background's influence. We also employ high-resolution feature preservation and incorporate a spatial-temporal attention module into the network to capture as many target features as possible and improve detection accuracy. Our method achieves superior results on the infrared image weak aircraft target detection dataset proposed by the National University of Defense Technology, as well as on the simulated dataset generated based on real-world observation. This demonstrates the efficiency of our approach for detecting weak point targets in infrared continuous images.

Peer Review Statement

All papers published in this volume have been reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.1. Type of peer review: Double anonymous2. Conference submission management system: Morressier3. Number of submissions received: 1784. Number of submissions sent for review: 1335. Number of submissions accepted: 1026. Acceptance Rate (Submissions Accepted / Submissions Received × 100): 57.3%7. Average number of reviews per paper: 2.008. Total number of reviewers involved: 659. Contact person for queries:Name: Wen ZhangAffiliation: National University of Defense Technology, ChinaEmail: zhangwen06@nudt.edu.cn

Preface

2023 International Conference on Frontiers of Ocean Science and Technology (ICFOST), sponsored by the College of Meteorology and Oceanology, National University of Defense Technology, was held successfully in Jiusuo Hotel Changsha from October 13 to 15, 2023.The Organizing Committees have invited experts in Marine science, Marine technology, Marine engineering, Marine resource exploration and application, Marine equipment R&D and manufacturing, etc. for academic exchanges. The Conference includes discussions on topics such as ocean circulation, air-sea interaction and ocean climate, ocean mesoscale and small-scale processes, ocean remote sensing and data analysis, ocean prediction, ocean environmental safety and support, ocean exploration technology, artificial intelligence oceanography, underwater acoustics, underwater acoustics signal processing, and oceanography teaching skills with a view to promoting the common development of industries, universities and research institutes in the field of ocean science and technologies.This year, ICFOST 2023 will consist of 10 oral sessions, 4 keynote talks from Prof. Juncheng Wang (National Engineering and Technology Research Center for Ocean Monitoring Equipment, China), Prof. Wenju Cai (Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia), Prof. Fangli Qiao (First Institute of Oceanography, Ministry of Natural Resources, China), Prof. Lianglong Da (Naval Submarine Academy, China), as well as 45 invited talks given by distinguished scholars and researchers from the related academic field. It will be a golden opportunity for the students, researchers and engineers to interact among themselves to get better views on technical development and trends for the near future.On behalf of the organizing committees, we’d like to express our sincere gratitude to everyone who has contributed to ICFOST 2023 as its success could have only been achieved through a team effort. Our special thanks go to all the conference speakers for their insightful reviews on many emerging research topics. We would like to thank the program chairs and the session chairs, for their support to the conference; as well as all the technical committee members and reviewers for their excellent work in promoting the conference and reviewing the papers. Finally, we are grateful to all the authors and presenters of the papers as well as all the attendees for their contributions to this wonderful conference.Warmest regards,ICFOST 2023 Organizing CommitteesList of Conference Committees are available in this Pdf.

Transplantation and optimization of molecular dynamics simulation on MT-3000

Molecular dynamics (MD) simulation is a research method to simulate the change of particle systems with the help of the powerful calculating power of computers. As the number of atoms increases and the duration of the simulation system extends, current mainstream software for MD simulations has become increasingly strained. This is largely due to a surge in computing time, attributed to the architecture characteristics of the central processing unit (CPU), which executes instructions serially and accesses shared memory randomly. These features have resulted in an enormous overhead that these software applications can no longer efficiently manage. Recently, National University of Defense Technology proposed the MT-3000 processor, a multi-zone heterogeneous architecture with a double-precision peak performance of 11.5 TFLOPS. In this paper, we implement the heterogeneous transplantation and optimization of the massively parallel program for MD simulation on Tianhe new-generation supercomputer based on MT-3000. By porting the key computing hotspot module to the acceleration zones of MT-3000, we have greatly improved the efficiency and performance of the simulation of billion-metal-atom cascade collision. The experimental results show that the heterogeneous MD program on a single node can achieve about 6 times faster than the original version executed in a general purpose zone of MT-3000. On the basis of 250 thousand cores, the strong-scalability parallel efficiency of heterogeneous MD program using 4 million cores can reach 89.89%, and the weak-scalability parallel efficiency can reach 91.98%. It can be predicted that the heterogeneous MD implementation has excellent scaling ability to simulate large-scale atomic cascade collisions on ten million of cores.

Theoretical validation and experimental calibration of a variable Mach number tunnel with continuous wave-elimination

The increasing demand for wide-range aircraft techniques drives the development of variable Mach number test facilities. Existing tunnels can provide uniform flow at design case, but the imperfect wave-elimination at non-design cases leads to insufficient flow uniformity, limited variation range, and complex operation. By applying flowfield inverse design and elasticity inverse design, this paper designs a continuous wave-elimination variable Mach number nozzle. The nozzle is validated by the perfect nozzle theory. By a single jack, the nozzle can deform to the perfect nozzles with different Mach numbers, eliminating the Mach waves continuously and achieving the high-quality test flow. The numerical evaluation validates the design and supports the fabrication of continuous wave-elimination variable Mach number tunnel for the National University of Defense Technology. The flow visualizations by Schlieren and nanoparticle-tracer based planar laser scattering technique indicate no vortices or waves induced by imperfect nozzle geometry. The Mach number calibrations show that the maximum deviations are less than 0.6%, and the standard deviations meet the national advanced standards. These results show that the tunnel eliminates the Mach waves at different Mach numbers through simple operation, providing a wider variable Mach range and a higher flow quality.

Airborne Measurement of Earth Gravity by Comparing Three Kinds of Gravimeters

NUDT (National University of Defense Technology) has developed a strapdown gravimeter based on the SINS/GPS named SGA-WZ01. From April to May 2012, an airborne gravimetry flight test was carried out in the South China Sea. In this flight test, except SGA-WZ01, a 3-axis inertial platform gravimeter, GT-1A, and a damped 2-axis platform gravimeter, TAGS, were installed to obtain the gravity information. This was the first time that such a flight campaign was undertaken in China. The flight test contains one repeated line flight and four gridded line flights. The evaluation of the repeatability is based on repeated surveying lines, and the evaluation of the internal accuracy is based on all gridded lines. The results show that both GT-1A and TAGS have systematic biases about several mGal for repeated lines along west-east and northwest-southeast repeated lines, and the repeatability of SGA-WZ01 is consistent along repeated lines in different directions. The accuracy of gridded lines without crossover adjustment of GT-1A is about 1.73 mGal, while the crossover accuracy is 2.76 mGal and 2.08 mGal for TAGS and SGA-WZ01, respectively. The outcomes show that the performance of the SGA-WZ01 is comparable to that of foreign gravimeters.

Foreword: A special issue on mechanical engineering to celebrate the 70th anniversary of the founding of the National University of Defense Technology

Foreword: A special issue on mechanical engineering to celebrate the 70th anniversary of the founding of the National University of Defense Technology

Peer review declaration

All conference organisers/editors are required to declare details about their peer review. Therefore, please provide the following information:• Type of peer review: Double-blind • Conference submission management system: Iconf submission system • Number of submissions received: 202 papers • Number of submissions sent for review: 202 papers • Number of submissions accepted: 102 papers • Acceptance Rate (Number of Submissions Accepted / Number of Submissions Received X 100): 50.5% • Average number of reviews per paper: 2 reviews • Total number of reviewers involved: 119 reviewers • Any additional info on review process: 1. Preliminary Review2. The papers passed the first review will be reviewed again from the following aspects: Originality, Innovation, Technical Merit, Applicability, Clarity of presentation and Relevance Contact person for queries: Name: Wen ZhangAffiliation: National University of Defense Technology, ChinaEmail: zhangwen06@nudt.edu.cn

Preface to the special topic: The 70 th anniversary of National University of Defense Technology

Preface to the special topic: The 70 th anniversary of National University of Defense Technology

First unmanned aerial vehicle airborne gravimetry based on the CH-4 UAV in China

As one of the important ways to obtain gravity fields, airborne gravimetry has been widely used because of its high efficiency and multiple scenario applicability. Compared with the manned aircraft commonly used in airborne gravimetry, unmanned aerial vehicle (UAV) platforms will provide a safer environment for pilots and surveyors. Based on the airborne gravimetry system integrated within a CH-4 UAV, this study carried out multiple gravimetry sorties in the survey area, including long-endurance gravimetry sorties. The test results show that the SGA-WZ04 gravimeter developed by the National University of Defense Technology (NUDT) can carry out gravimetry tests for >20 h on the modified CH-4 UAV platform and complete measurement tasks safely and reliably. From analysis of the gravity data processing results, the repeatability of the repeat lines is 0.35 mGal/3 km, and the external coincidence accuracy after fitting with the EGM2008 model is 0.47 mGal/3 km. The results provide strong support for the future expansion and layout of airborne gravimetry in areas where the use of manned aircraft is costly and operationally inefficient, such as medium and distant seas.

Design and Experimental Validation of a Micro-Newton Torsional Thrust Balance for Ionic Liquid Electrospray Thruster

This work describes the advances on a micro-newton torsional thrust balance for ionic liquid electrospray thruster (ILET) being developed in the National University of Defense Technology. The torsional thrust balance adopts an asymmetric pendulum arm about a flexural pivot, and an electrostatic comb device is used for calibration, which makes the balance compact and allows the measurement of the micro-newton level thrust with high accuracy. To minimize the influence of gravity on the measurement results, a two-dimensional adjustable counterweight mechanism is adapted to balance the entire arm and make its mass center close to the pivot axis. Mechanical oscillations are passively damped with an eddy current damper. A series of experimental studies are carried out in a vacuum chamber; the results provide validation that the balance has good linearity in the range of 2–30 μN, with the resolution better than 0.21 μN and the settling time to a step force is less than 7 s, which can meet the thrust measurement requirements of the ILET operating in the mode of alternating work voltage polarity. In addition, we find that the electrostatic force generated by the power supply wire has a non-negligible influence on thrust measurement results, which needs to be considered in the actual thrust measurement.

Dyson sphere building: On the design of the GTOC11 problem and summary of the results

The National University of Defense Technology and the Xi’an Satellite Control Center organized the 11th edition of the Global Trajectory Optimization Competition (GTOC11) in 2021. The GTOC11 problem was created as a link between the ninth and tenth editions of the Global Trajectory Optimization Competition to bridge the gap between the planetary and galactic civilizations by introducing an intermediate stellar civilization scenario. The problem involves the construction of a Dyson sphere, a theoretical mega-structure that encircles a star with platforms orbiting in a tight formation to capture the maximum energy from it. Challenges in astrodynamics including the construction-orbit selection, combinatorial flyby of multiple asteroids, and mass distribution among stations were considered in the Dyson sphere design. A total of 94 teams registered for the competition, of which 25 teams provided solutions and passed automatic verification on the website. In this article, we describe the selection of the problem and its design process. In addition, an overview of the entire competition and an analysis of its results are presented.

Preface

It is our great pleasure to welcome you to the 2021 12th Asia Conference on Mechanical and Aerospace Engineering (ACMAE) on December 29-31th, 2021, which is planned to hold in Nanjing University of Science and Technology, China. However, Since the outbreak of the global covid-19 pandemic, it has been a challenge for the academic community to organize international conferences due to travel restrictions. Considering the safety of the participants. This time we have decided to take the conference with combination of online and offline. ACMAE 2021 is dedicated to issues related to the advancement of all Mechanical and Aerospace Engineering topics.Mechanical Engineering of Nanjing University of Science and Technology, as the host of ACMAE 2021 attaches great importance to international academic exchanges and joint talent training, and has established cooperative relations with Baumann Moscow State Technical University, Samara State University and etc. Series of conferences have been organized as International Conference on Mechanical and Electrical Systems and Testing Technology, International Forum on Special Equipment and Engineering Mechanics, etc.We have honor to invite Prof. K.A. Neusypin and Prof. Chen Xiong served as the chairman of the conference respectively. In recent years, ACMAE has been successfully organized online in Yokohama, Japan (2017), Singapore (2018), Bangkok, Thailand (2019), and 2020, with the support of Japan National Defense University, and Chulalongkorn University and etc. respectively. The conference aims to build an annual academic exchange platform, focusing on the scientific research and discussion in the field of aerospace in Asia, promoting the exchange and cooperation between Asian scholars, and showing the booming development of aerospace in Asia to the world.The conference invited two top international experts and three top domestic experts in the field of aerospace to deliver keynote speeches, including Professor Genxiang Xia from university of Delaware and IEEE Member. Professor Qu Ronghai, member of degree Committee, Huazhong University of Science and Technology, Professor Luling An, expert of Nanjing University of Aeronautics and Astronautics; Prof. Wu Weitao, Professor of Nanjing University of Science and Technology, member of the Committee of Physics and Aerodynamics of Chinese Society of Aerodynamics; Professor Maria Selezneva, Deputy Head of The Department of Instrument Engineering and Technology, Bauman State Technical University, Moscow, Russia. Each speaker has delivered the speech more than 30 minutes. In addition, more than 100 experts and scholars from the world's top research institutes of aerospace technology, including Samara University in Russia, as well as domestic senior research institutes of aerospace technology, including Nanjing University of Aeronautics and Astronautics, Northwestern Polytechnical University, National University of Defense Technology, Shanghai Jiao Tong University and Xiamen University, attended the conference.This conference proceeding, contains a collection of research papers, organized as chapters, presented at ACMAE 2021. This publication, containing the conference proceedings, documents and presentations made at ACMAE 2021, is the result of the creative work of their authors and of a highly selective peer-review process.We are also grateful to colleagues for committing their active participation and promoting the conference during this difficult time. In this specific time, although we are not able to meet, I believe online conferences like these would shorten the distance between people, reduce spending and inconvenience on the way. Wish to see all of you in next year conference.Conference ChairProf. Xiong ChenACMAE 2021List of titles Organizing Committees are available in this Pdf.

The Development and Innovation of the First Supercomputer YH-1 in China: From the Perspective of Technology History

China’s first supercomputer capable of 100 million calculations per second was the YH-1, which was independently developed by the Institute of Computer Science at the National University of Defense Technology (NUDT) between 1978 and 1983. YH-1 played an important role in China’s national defense construction and national economic development. It made China one of the few countries in the world to successfully develop a supercomputer. Based on original archive documents, interviews with relevant personnel, and an analysis of the technological parameters of the supercomputers YH-1 in China and Cray-1 in the United States, this paper reviews in detail the historic process of the development of YH-1, analyzing its innovation and summarizing the experience and lessons learned from it. This analysis is significant for current military-civilian integration, and the commercialization of university research findings in China.

Precise Relative Orbit Determination for Chinese TH-2 Satellite Formation Using Onboard GPS and BDS2 Observations

TH-2 is China’s first short-range satellite formation system used to realize interferometric synthetic aperture radar (InSAR) technology. In order to achieve the mission goal of InSAR processing, the relative orbit must be determined with high accuracy. In this study, the precise relative orbit determination (PROD) for TH-2 based on global positioning system (GPS), second-generation BeiDou navagation satellite system (BDS2), and GPS + BDS2 observations was performed. First, the performance of onboard GPS and BDS2 measurements were assessed by analyzing the available data, code multipath errors and noise levels of carrier phase observations. The differences between the National University of Defense Technology (NDT) and the Xi’an Research Institute of Surveying and Mapping (CHS) baseline solutions exhibited an RMS of 1.48 mm outside maneuver periods. The GPS-based orbit was used as a reference orbit to evaluate the BDS2-based orbit and the GPS + BDS2-based orbit. It is the first time BDS2 has been applied to the PROD of low Earth orbit (LEO) satellite formation. The results showed that the root mean square (RMS) of difference between the PROD results using GPS and BDS2 measurements in 3D components was 2.89 mm in the Asia-Pacific region. We assigned different weights to geostationary Earth orbit (GEO) satellites to illustrate the impact of GEO satellites on PROD, and the accuracy of PROD was improved to 7.08 mm with the GEO weighting strategy. Finally, relative orbits were derived from the combined GPS and BDS2 data. When BDS2 was added on the basis of GPS, the average number of visible navigation satellites from TH-2A and TH-2B improved from 7.5 to 9.5. The RMS of the difference between the GPS + BDS2-based orbit and the GPS-based orbit was about 1.2 mm in 3D. The overlap comparison results showed that the combined orbit consistencies were below 1 mm in the radial (R), along-track (T), and cross-track (N) directions. Furthermore, when BDS2 co-worked with GPS, the average of the ambiguity dilution of precision (ADOP) reduced from 0.160 cycle to 0.153 cycle, which was about a 4.4% reduction. The experimental results indicate that millimeter-level PROD results for TH-2 satellite formation can be obtained by using onboard GPS and BDS2 observations, and multi-GNSS can further improve the accuracy and reliability of PROD.

Theoretical calculations of hyperfine splitting, Zeeman shifts, and isotope shifts of 27Al+ and logical ions in Al+ clocks* *Project supported by the National Natural Science Foundation of China (Grant Nos. 11604385 and 91536106), the Natural Science Foundation of Hunan Province, China (Grant No. 2019JJ50743), and the Research Project of the National University of Defense Technology (Grant No. ZK17-03-11).

Based on the multiconfiguration Dirac-Hartree-Fock (MCDHF) method, similar models are employed to simultaneously calculate the first-order and second-order Zeeman coefficients as well as the hyperfine interaction constants of the related energy levels of 27Al+ and its logical ions 9Be+ and 25Mg+ in the 27AI+ optical clock. With less than 0.34% deviations from experimental values in Zeeman coefficients of 27Al+, these calculated parameters will be of great help for better evaluation of the systematic uncertainty. We also calculate the isotope shift parameters of the related energy levels, which could extend our knowledge and understanding of nuclear properties of these ions.

Optimization of the parallel semi‐Lagrangian scheme based on overlapping communication with computation in the YHGSM

AbstractThe Yi‐he global spectral model (YHGSM) is a numerical weather prediction (NWP) model developed by the National University of Defense Technology, China. The YHGSM has two optimization techniques that have recently been adopted for the on‐demand communication version of the semi‐Lagrangian scheme: one‐sided communication and an adaptive strategy for the maximum wind speed. However, the limitation of the original model is that the computation and communication cannot be performed simultaneously, resulting in a waste of resources. This article addresses this limitation by dividing the array to be interpolated into three independent subarrays. The communication and interpolation operations for each subarray are rearranged into three groups. This allows the overlap of the communication with computation between different groups. In addition, by adjusting the synchronization operations, the degree of overlap is further increased. In the optimized scheme, the per‐step overhead of the interpolation phase is greatly reduced, and its speedup decreases first and increases afterwards with the increase of the number of cores. Therefore, when the number of cores is large, the benefits from overlapping the communication with computation will be more substantial.

Advancing DSP into HPC, AI, and beyond: challenges, mechanisms, and future directions

Digital Signal Processors (DSPs) have been widely used in embedded domains, delivering high performance with ultra-low power consumption. Such promises make it attractive for more domains that DSP was not an option before. To show how DSP lives up to these promises, we review two milestone DSPs: FT-Matrix and FT-Matrix2, which are designed by National University of Defense Technology with the purpose of advancing DSPs into the era of higher performance computing, AI, and even beyond. We demonstrate that the key challenges lie in the orchestration of huge computation resources and efficient data supply sub-system design. We show the key mechanisms in both FT-Matrix and FT-Matrix2 targeting these challenges, and also come up with possible future directions for enabling DSPs for a wider scope of applications.

Novel energy dissipative method on the adaptive spatial discretization for the Allen–Cahn equation* *Project supported by the National Key R&D Program of China (Grant No. 2020YFA0709800), the National Natural Science Foundation of China (Grant Nos. 11901577, 11971481, 12071481, and 12001539), the Natural Science Foundation of Hunan, China (Grant Nos. S2017JJQNJJ0764 and 2020JJ5652), the fund from Hunan Provincial Key Laboratory of

We propose a novel energy dissipative method for the Allen–Cahn equation on nonuniform grids. For spatial discretization, the classical central difference method is utilized, while the average vector field method is applied for time discretization. Compared with the average vector field method on the uniform mesh, the proposed method can involve fewer grid points and achieve better numerical performance over long time simulation. This is due to the moving mesh method, which can concentrate the grid points more densely where the solution changes drastically. Numerical experiments are provided to illustrate the advantages of the proposed concrete adaptive energy dissipative scheme under large time and space steps over a long time.