Aerial Object Research Articles

The Methods for Improving the Quality of Detection of Inconspicuous Aerial Objects Through the Use of External Radiation Sources

The article focuses on the methods for improving the quality of detection of inconspicuous aerial objects through the use of external radiation sources. It has been established that the existing methods for improving the quality of detection of airborne objects do not meet modern requirements for detection indicators. The energy potential of modern radars is limited. Therefore, the article proposes detection methods that combine active radar and nontraditional principles for detecting airborne objects. In the article, we will focus only on two main methods for improving the quality of detection of small airborne objects. This is the use of energy from digital television receivers (external radiation source) and the use of information from ADS-B receivers. Technically, this means introducing an additional receiving channel into the radar. The gain in the conditional probability of correct detection for bistatic signal reception (compared to monostatic) is due to the properties of the bistatic effective scattering surface. It is shown that with the additional use of digital television signals, the quality of detection of airborne objects is improved. It is shown that the additional use of information from the ADS-B receivers makes it possible to increase the accuracy of determining the coordinates of an air object up to 40-50%, depending on the parameters of the movement of an air object. The subject of advanced research is the development of methods for suppressing the penetrating signal when using digital television and the development of methods for detecting airborne objects not equipped with the ADS B system. Keywords— aerial object, detection, quality, third-party source, ADS-B, digital TV, radar

Multi-Band Acoustic Monitoring of Aerial Signatures

The acoustic monitoring, omni-directional system (AMOS) in the Galileo Project is a passive, multi-band, field microphone suite designed to aid in the detection and characterization of aerial phenomena. Acoustic monitoring augments the Project’s electromagnetic sensors suite by providing a relatively independent physical signal modality with which to validate the identification of known phenomena and to more fully characterize detected objects. The AMOS system spans infrasonic frequencies down to 0.05[Formula: see text]Hz, all of audible, and ultrasonic frequencies up to 190[Formula: see text]kHz. It uses three distinct systems with overlapping bandwidths: infrasonic (0.05[Formula: see text]Hz – 20[Formula: see text]Hz), audible (10[Formula: see text]Hz – 20[Formula: see text]kHz), and ultrasonic (16[Formula: see text]kHz – 190[Formula: see text]kHz). The sensors and their capture devices allow AMOS to monitor and characterize the tremendous range of sounds produced by natural and human-made aerial phenomena, and to encompass possible acoustic characteristics of novel sources. Sound signals from aerial objects can be captured and classified with a single microphone under the following conditions: the sound reaches the sensor; the sound level is above ambient noise; and the signal has not been excessively distorted by the transmission path. A preliminary examination of the signal and noise environment required for the detection and characterization of aerial objects, based on theoretical and empirical equations for sound attenuation in air, finds that moderately loud audible sources (100[Formula: see text]dB) at 1[Formula: see text]km are detectable, especially for frequencies below 1[Formula: see text]kHz and in quiet, rural environments. Infrasonic sources are detectable at much longer distances and ultrasonic at much shorter distances. Preliminary aircraft recordings collected using the single, omni-directional audible microphone are presented, along with basic spectral analysis. Such data will be used in conjunction with flight transponder data to develop algorithms and corresponding software for quickly identifying known aircraft and characterizing the sound transmission path. Future work will include multi-sensor audible and infrasonic arrays for sound localization; analysis of larger and more diverse data sets; and exploration of machine learning and artificial intelligence integration for the detection and identification of many more types of known phenomena in all three frequency bands.

SkyWatch: A Passive Multistatic Radar Network for the Measurement of Object Position and Velocity

Quantitative three-dimensional (3D) position and velocity estimates obtained by passive radar will assist the Galileo Project in the detection and classification of aerial objects by providing critical measurements of range, location, and kinematics. These parameters will be combined with those derived from the Project’s suite of electromagnetic sensors and used to separate known aerial objects from those exhibiting anomalous kinematics. SkyWatch, a passive multistatic radar system based on commercial broadcast FM radio transmitters of opportunity, is a network of receivers spaced at geographical scales that enables estimation of the 3D position and velocity time series of objects at altitudes up to 80[Formula: see text]km, horizontal distances up to 150[Formula: see text]km, and at velocities to [Formula: see text][Formula: see text]km/s ([Formula: see text][Formula: see text]Mach). The receivers are designed to collect useful data in a variety of environments varying by terrain, transmitter power, relative transmitter distance, adjacent channel strength, etc. In some cases, the direct signal from the transmitter may be large enough to be used as the reference with which the echoes are correlated. In other cases, the direct signal may be weak or absent, in which case a reference is communicated to the receiver from another network node via the internet for echo correlation. Various techniques are discussed specific to the two modes of operation and a hybrid mode. Delay and Doppler data are sent via internet to a central server where triangulation is used to deduce time series of 3D positions and velocities. A multiple receiver (multistatic) radar experiment is undergoing Phase 1 testing, with several receivers placed at various distances around the Harvard–Smithsonian Center for Astrophysics (CfA), to validate full 3D position and velocity recovery. The experimental multistatic system intermittently records raw data for later processing to aid development. The results of the multistatic experiment will inform the design of a compact, economical receiver intended for deployment in a large-scale, mass-deployed mesh network. Such a network would greatly increase the probability of detecting and recording the movements of aerial objects with anomalous kinematics suggestive of Unidentified Aerial Phenomena (UAP).

A Hardware and Software Platform for Aerial Object Localization

To date, there are little reliable data on the position, velocity and acceleration characteristics of Unidentified Aerial Phenomena (UAP). The dual hardware and software system described in this document provides a means to address this gap. We describe a weatherized multi-camera system which can capture images in the visible, infrared and near infrared wavelengths. We then describe the software we will use to calibrate the cameras and to robustly localize objects-of-interest in three dimensions. We show how object localizations captured over time will be used to compute the velocity and acceleration of airborne objects.

Free-Form Deformation Parameterization on the Aerodynamic Optimization of Morphing Trailing Edge

Every aerodynamic optimization is proceeded by a parameterization of the studied aerial object, and due to its influence on the final optimization process, careful attention should be made in choosing the appropriate parameterization method. An aerodynamic optimization of a morphing trailing edge is performed using a free-form deformation parameterization technique with the purpose of examining the influence of the initial conditions of the parameterization on the optimization results, namely on the number of control points. High-fidelity gradient-based optimization using the discrete adjoint method is established by the coupling of OpenFOAM and Python within the DAFoam optimization framework. The results indicate that the number of control points has a considerable effect on the optimization process, in particular on the convergence, objective function value, and on the deformation feasibility.

Разработка имитационной модели определения координат воздушных объектов по отраженному от воздушного объекта и принятому на антенну навигационного приемника навигационному сигналу ГНСС

This article is devoted to the development of a simulation model for determining the coordinates of air objects based on the GNSS navigation signal reflected from the air object and received on the antenna of the navigation receiver. To solve the problem of radar of aerial objects using bistatic location based on the use of GNSS satellites, it was necessary to pre-evaluate the energy or energy potential of the satellite navigation signal reflected from the aerial object, which then comes to the receiving antenna of the navigation equipment of consumers. The results obtained in the form of the developed computational schemes will form the basis of the algorithmic and software of the passive radar system and will allow determining the requirements for its technical support carried out within the framework of the grant project of the Ministry of Education and Science of the Republic of Kazakhstan No. AP09260581. In this paper, simulation modeling is understood as such an approach to modeling complex systems, which allows us to develop a single meta-model that combines into an integral set of models of objects and processes of their interaction based on a single information and event space in a single discrete time scale, and implement the resulting meta-model in an electronically equivalent form on an electronic computer. One of the main tasks of the study of complex radio engineering systems by modeling is the choice of modeling principles that should allow the construction of the simulated system in accordance with the goals and objectives of the study in such a way that the model adequately reflects the processes and patterns of behavior of the system under study with the necessary degree of reliability and detail. One of the most productive methods of developing a software and hardware complex of simulation modeling is object-oriented modeling.

WSODet: A Weakly Supervised Oriented Detector for Aerial Object Detection

In contrast to natural objects, aerial targets are usually non-axis aligned with arbitrary orientations. However, mainstream Weakly Supervised Object Detection (WSOD) methods can only predict horizontal bounding boxes (HBB) from existing proposals generated by offline algorithms. To predict oriented bounding boxes (OBB) for aerial targets while testing images end-to-end without proposals, WSODet is designed leveraging on Layer-wise Relevance Propagation (LRP) and Point Set Representation (RepPoints). To be specific, based on the mainstream WSOD framework, Layer-wise Relevance Propagation on Multiple Instance Learning branch (MIL-LRP) is conducted to decrease the uncertainty and ambiguity of feature map. Then, a pseudo oriented label generation algorithm is designed to obtain OBB pseudo labels, which serve as supervision to train an Oriented RepPoint Net under the guidance of improved oriented loss function. During the test, input images are sent to oriented RepPoint branch to obtain OBB predictions without proposals. Extensive experiments on DIOR, NWPU VHR-10.v2 and HRSC2016 datasets demonstrate the effectiveness of our method to predict precise oriented aerial objects, achieving 22.2%, 46.5% and 43.3% mAP, respectively. Moreover, training jointly with oriented RepPoint branch boosts the results of original WSOD framework compared with existing WSOD methods even if there is no specific design for original structure.

MUS-CDB: Mixed Uncertainty Sampling With Class Distribution Balancing for Active Annotation in Aerial Object Detection

Recent aerial object detection models rely on a large amount of labeled training data, which requires unaffordable manual labeling costs in large aerial scenes with dense objects. Active learning effectively reduces the data labeling cost by selectively querying the informative and representative unlabelled samples. However, existing active learning methods are mainly with class-balanced settings and image-based querying for generic object detection tasks, which are less applicable to aerial object detection scenarios due to the long-tailed class distribution and dense small objects in aerial scenes. In this paper, we propose a novel active learning method for cost-effective aerial object detection. Specifically, both object-level and image-level informativeness are considered in the object selection to refrain from redundant and myopic querying. Besides, an easy-to-use class-balancing criterion is incorporated to favor the minority objects to alleviate the long-tailed class distribution problem in model training. We further devise a training loss to mine the latent knowledge in the unlabeled image regions. Extensive experiments are conducted on the DOTA-v1.0 and DOTA-v2.0 benchmarks to validate the effectiveness of the proposed method. For the ReDet, KLD, and SASM detectors on the DOTA-v2.0 dataset, the results show that our proposed MUS-CDB method can save nearly 75% of the labeling cost while achieving comparable performance to other active learning methods in terms of mAP. Code is publicly online.

Toward Hierarchical Adaptive Alignment for Aerial Object Detection in Remote Sensing Images

Toward Hierarchical Adaptive Alignment for Aerial Object Detection in Remote Sensing Images

Learnable Loss Balancing in Anchor-Free Oriented Detectors for Aerial Object

Oriented object detection plays an important role in aerial image interpretation. Image processing speed is also essential due to massive amounts of aerial images. Anchor-free oriented detectors with fast processing speed are generally accepted despite the absence of pre-set anchors, contributing to their performance gap with anchor-based detectors. Most anchor-free oriented detectors are carefully designed by defining samples according to target characteristics, which require substantial prior knowledge, to realize improved performance. This study proposes an anchor-free oriented detector (termed as rfpoint) that requires minimal prior knowledge. Moreover, this study mainly aims to introduce a dynamic sample definition strategy. This strategy is modeled as a dynamic regulating process, wherein the classification and box regression interact until the model converges. A rotating quality-driven loss (RQDL) and adaptive-weight box loss (AWBL) are also proposed to realize the aforementioned process. RQDL redefines positive and negative attributes of samples according to the distribution of rotating Intersection-over-Unit (IoU) between predictions and ground truth. AWBL adjusts the importance degree of candidate samples in the box regression through classification scores. The proposed method is then tested on three mainstream aerial image datasets (DOTA, DIOR, and HRSC2016). Results reveal that the proposed method achieves the best performance compared with other oriented detectors, whose mAP are 79.92%, 70.88%, and 90.67%. Moreover, the method maintains the inference speed advantage of anchor-free detectors. An effective sample definition method can bridge the performance gap of anchor-free oriented detectors without minimizing inference speed.

Detection characteristics and features of signal processing of aerial objects on images of spacborne synthetic aperture radars

The article considers the possibilities of detecting airborne moving targets by space radars with aperture synthesis using interferometry mode along the path. The method of calculating the signal-to-noise ratio, the characteristics of detecting aerial targets in the mode of selection of moving targets is described. Based on calculations, it is shown that the dynamics of the underlying surface caused by fluctuations in grass, foliage or an agitated sea surface significantly reduces the ability to detect aerial objects by space radars with aperture synthesis in the selection of moving targets mode. It is shown by the examples of existing space radars with aperture synthesis that when using the interferometry method along the path, at least 2 interferometric bases and refocusing on the marks of bright moving targets in the mode of selection of moving targets channel, the effectiveness of the detection procedure can be sufficient for practical applications. However, the region of small radial velocity values remains a blind zone for any multi-base circuit configurations and can only be reduced by increasing the energy potential of the space radars with aperture synthesis.

Comparative analysis of interference protection of "Friend-Foe" radar identification systems

The paper examines the existing systems of radar identification based on the "friend-foe" feature from the point of view of immunity evaluation. Currently, there are two systems of the radar identification based on the characteristic of "friend-foe", namely, the "Password" and MkXIIA. They are the radar identification systems based on the "home-foreign" feature. The former of the specified systems operates in a frequency range that differs from the frequency range of secondary radar systems, while the second one operates in the frequency range of secondary radar. These systems are ones of the main information resources of the airspace control system and are built on the principles of a one-channel or two-channel information transmission system. It allows the interested party, both an unauthorized use of this information resource for remote determination of the coordinates of aerial objects, on the one hand, and twisting the information of these information resources, on the other hand, which leads to unpredictable results. The purpose of the work is to assess the immunity of existing radar identification systems based on the "friend-foe" feature. The analysis of the interference protection of existing systems for the object radar identification based on the "friend-foe" feature, built on the principles of interrogative and non-interrogative information systems presented in this work, showed that the use of rectangular radio signals with time-pulse modulation as request and response signals emitting by air objects, has low immunity and excludes the energy stealthiest of the respondents of aerial objects. And, as a result, it allows for unauthorized calculation of the coordinates of air objects by the interested party based on the emitted identification signals on the basis of "friend-foe".

Friend-or-Foe Recognition Algorithm Development for the Corresponding Software Building

The year 2022 showed an urgent need to improve the existing systems for recognizing objects in the aerial space, which is caused by the significant increase in the number of technical means (especially unmanned aerial vehicles) on the battlefield. Such a sharp increase in the number of objects that simultaneously take part in combat operations in the air requires the improvement of military object recognition systems, both qualitatively and quantitatively. This requires the development of appropriate new generation Friend-or-Foe algorithms for the objects’ recognition. The main requirements for recognition systems of aerial objects of civil application were determined. They includes: maximum com- patibility; support for a large number of objects; outdated recognition complexes support; support for alternative ways of recognition; support for alternative data entry methods; determining the coordinates of aerial objects in an emergency situation. Friend-or-foe recognition systems for military applications are also considered. In contrast to civilian systems, the following basic requirements have been identified for them: 1) Maximum speed of the recognition process. 2) Protection against false positives. 3) Protection against legitimate aerial object imitation. 4) Support for a large number of objects. 5) Protection against cases of loss of a legitimate aerial object. 6) Rotation of the secret part. 7) Protection against false-negative results to prevent friendly fire. 8) Protection against man-in-the-middle attacks. 9) Flexible integration with the NATO block recognition system. 10) Availability of opportunities for purely domestic production and support of the object recognition system. 11) Protection against electronic warfare means. 12) Support for several recognition modes. 13) Automatic blocking of the launch of ground-to-air and air-to-air weapons against objects that confirm their legitimacy by a correct response to a request. 14) Determining the coordinates of aerial objects in an emergency. Based on the formulated requirements, a new friend-or-foe algorithm for the state identification system for military use is proposed, built based on the state standards, and taking into account the features of its software implementation in order to increase speed. Its implementation will ensure sufficient scalability, stability, reliability, and multi-level recognition.

Sparse Channel Pruning and Assistant Distillation for Faster Aerial Object Detection

In recent years, object detectors based on convolutional neural networks have been widely used on remote sensing images. However, the improvement of their detection performance depends on a deeper convolution layer and a complex convolution structure, resulting in a significant increase in the storage space and computational complexity. Although previous works have designed a variety of new lightweight convolution and compression algorithms, these works often require complex manual design and cause the detector to be greatly modified, which makes it difficult to directly apply the algorithms to different detectors and general hardware. Therefore, this paper proposes an iterative pruning framework based on assistant distillation. Specifically, a structured sparse pruning strategy for detectors is proposed. By taking the channel scaling factor as a representation of the weight importance, the channels of the network are pruned and the detector is greatly slimmed. Then, a teacher assistant distillation model is proposed to recover the network performance after compression. The intermediate models retained in the pruning process are used as assistant models. By way of the teachers distilling the assistants and the assistants distilling the students, the students’ underfitting caused by the difference in capacity between teachers and students is eliminated, thus effectively restoring the network performance. By using this compression framework, we can greatly compress the network without changing the network structure and can obtain the support of any hardware platform and deep learning library. Extensive experiments show that compared with existing detection networks, our method can achieve an effective balance between speed and accuracy on three commonly used remote sensing target datasets (i.e., NWPU VHR-10, RSOD, and DOTA).

Threshold sensitivity of staring thermal imaging devices operating in slant atmospheric paths

Subject of study. A method for predicting threshold sensitivity, i.e., the resulting noise equivalent temperature difference of modern air- or ground-based staring thermal imaging devices, was considered. Aim of study. The study aimed to estimate the effectiveness of thermal imaging devices surveying aerial and ground-based objects. Method. The method implied a theoretical computational analysis based on a previously developed model for emission of the atmosphere in arbitrarily oriented paths considering the following: the nominal noise equivalent temperature difference that characterizes the quantum or thermal array photodetectors used in thermal imaging devices; the temperature profile of the atmosphere whose emission, together with background emission of, e.g., the ground surface, significantly affects the external photon noise in thermal imaging devices; elevation of the line of sight of the object; presence and type of clouds; and season. Results. A method was developed for refined calculation of the threshold sensitivity of staring thermal imaging devices determined by photon noise resulting from the radiation of the elements of the thermal imaging device itself, particularly its objective, dark-current noise, readout noise, and spatial (geometric) noise resulting from the residual dispersion of sensitivity of photodetector elements after correction. This threshold sensitivity determines the effectiveness of thermal imaging devices in object detection and recognition against a quasi-uniform background of ground or sky under different conditions, including long slant ground-to-air and air-to-ground atmospheric paths. The investigation results are presented in a form convenient for practical applications, i.e., reduced to equations and illustrated with a calculation example. Practical significance. The engineering method for calculating the resulting noise equivalent temperature difference proposed in this study, which considers the thermal stratification of the atmosphere and corresponds to real operating conditions of thermal imaging devices, is advantageous for adequate assessment of their effectiveness in surveilling ground-based and aerial objects, including low-contrast objects.

DEVELOPMENT OF A VIDEO PROCESSING MODULE FOR THE TASK OF AIR OBJECT RECOGNITION BASED ON THEIR CONTOURS

The subject of research in the article is the module of automatic segmentation and subtraction of the background, which is created, based on the sequential application of methods of image preprocessing and modified method of interactive segmentation of images and implemented in the system of optical monitoring of the air situation. The aim of the work is to develop an image segmentation module to increase the efficiency of recognition of an air object type on a video image in the system of visual monitoring of the air environment by means of qualitative automatic segmentation. To solve this problem, a modified interactive algorithm in the mode of automatic selection of an object in the image, which allows more accurately, without the participation of the operator, to determine the foreground pixels of the image for further recognition of the type of airborne object. The following tasks are solved in the article: the analysis of existing methods of binarization of color images for semantic segmentation of images, which are used in image recognition systems; the development of a pipeline of methods for automatic segmentation of images in the system of optical monitoring of the air environment. In the work, the following methods are used: methods of digital image processing, methods of filtering and semantic segmentation of images, methods of graph analysis. The following results are obtained: the results of image processing with the proposed module of segmentation and background subtraction confirm the performance of the module procedures. The developed pipeline of methods included in the module demonstrates correct segmentation in 93% of test images in automatic mode without operator participation, which allows us to conclude about the effectiveness of the proposed module. Conclusions: The implementation of the developed module of segmentation and background subtraction for the system of optical monitoring of the air environment allowed to solve the problem of segmentation of video images for further recognition of aerial objects in the system of optical monitoring of the air environment in automatic mode with a high degree of reliability, thus increasing the operational efficiency of this system.

Processing of information from networks of airspace surveillance radar systems

The presented paper examines the principles and methods for processing of information from airspace surveillance radar networks. Information technologies allow implementing automatic collection, processing, storage, transmission and issuance of radar information to users. In this work, the synthesis and analysis of the optimal structure of the interstage processing of signal data and radar information of primary processing in the network of airspace surveillance radar systems is carried out. The quality of information from radar networks of airspace surveillance systems affects almost all indicators of the quality of the radar systems network functioning. The stages of radar information processing in radar systems networks are analyzed. The importance of specifying the above stages for creating a complete picture of the air situation in the area of responsibility is shown. To improve the quality of information support for consumers, a network of radar systems requires information processing at all stages. At each stage of information processing, information processing quality indicators were analyzed. This made it possible to show that the staged implementation of information processing, on the one hand, simplified the optimization of processing within each processing stage, but on the other hand, it made it difficult to carry out compatible optimization of both the detection of an aerial object and the measurement of the coordinates of an aerial object. The synthesized structure of processing radar information of the network of radar systems of airspace surveillance, which in its turn made it possible to carry out interstage optimization of processing of both signal data and primary processing information. Calculations have shown that the method of information processing, in which the combination of information is carried out at the level of decision-making on the detection of airborne objects in each signal data processing channel, has some advantages in the quality of information processing of the network of radar systems compared to the one currently used. time option combination information at the stage of signal processing. At the same time, for the method of combining information at the level of decision-making about the detection of aerial objects, the flow of transmitted information to the joint processing point is significantly reduced. All this allows improving the quality of information processing in the airspace control system.

A Novel Multi-Scale Transformer for Object Detection in Aerial Scenes

Deep learning has promoted the research of object detection in aerial scenes. However, most of the existing networks are limited by the large-scale variation of objects and the confusion of category features. To overcome these limitations, this paper proposes a novel aerial object detection framework called DFCformer. DFCformer is mainly composed of three parts: the backbone network DMViT, which introduces deformation patch embedding and multi-scale adaptive self-attention to capture sufficient features of the objects; FRGC guides feature interaction layer by layer to break the barriers between feature layers and improve the information discrimination and processing ability of multi-scale critical features; CAIM adopts an attention mechanism to fuse multi-scale features to perform hierarchical reasoning on the relationship between different levels and fully utilize the complementary information in multi-scale features. Extensive experiments have been conducted on the FAIR1M dataset, and DFCformer shows its advantages by achieving the highest scores with stronger scene adaptability.

Unmanned aerial vehicles: potential for use in the storage systems of industrial companies

In the article, the authors investigate the problem of using unmanned aerial vehicles (drones) in company warehouses and consider their use as one of the ways to optimize the logistics business processes of companies. Technologies are considered, on the basis of which modern unmanned aerial vehicles, drone models of various companies, the main functionality of drone’s work. The authors high-light the strengths and weaknesses of unmanned aerial vehicles, key factors determining the success of the use of drones, and favorable conditions for their use in warehouses. The authors are considering the possibility of sharing drones and RFID technology. Barriers to the active introduction of drones into the logistics business processes of warehousing systems and requirements for warehouse complexes planning the use of unmanned aerial objects have been identified. Promising areas of application of technology in warehouse logistics are also considered: creation of anti-theft and fire protection systems, using drones to read CIS marks. The authors used data from analytical reports and data collected by the authors themselves, as well as empirical research methods.

Detection of plane in remote sensing images using super-resolution.

The object detection of remote sensing image often has low accuracy and high missed or false detection rate due to the large number of small objects, instance level noise and cloud occlusion. In this paper, a new object detection model based on SRGAN and YOLOV3 is proposed, which is called SR-YOLO. It solves the problems of SRGAN network sensitivity to hyper-parameters and modal collapse. Meanwhile, The FPN network in YOLOv3 is replaced by PANet, shortened the distance between the lowest and the highest layers, and the SR-YOLO model has strong robustness and high detection ability by using the enhanced path to enrich the characteristics of each layer. The experimental results on the UCAS-High Resolution Aerial Object Detection Dataset showed SR-YOLO has achieved excellent performance. Compared with YOLOv3, the average precision (AP) of SR-YOLO increased from 92.35% to 96.13%, the log-average miss rate (MR-2) decreased from 22% to 14%, and the Recall rate increased from 91.36% to 95.12%.