Aerial Object Research Articles

Boosting Point Set-Based Network with Optimal Transport Optimization for Oriented Object Detection

When handling complex remote sensing scenarios, rotational angle information can improve detection accuracy and enhance algorithm robustness, providing support for fine-grained detection. Point set representation is one of the most commonly used methods in arbitrary-oriented object detection tasks, leveraging discrete feature points to represent oriented targets and achieve high accuracy in angle prediction. However, due to the inherent discreteness of point set representation, it is prone to significant impact from isolated points and representational ambiguity in harsh application scenarios, leading to inaccurate detection. To address this issue, an efficient aerial object detector named BE-Det is proposed, which uses the optimal transport (OT) strategy to constrain the positions of isolated points. Additionally, a candidate point set quality evaluation scheme is designed to effectively assess the quality of candidate point sets. Experimental results on two challenging aerial datasets demonstrate that the proposed method outperforms several advanced detection methods.

Devising a method for detecting an aerial object by radar with an additional channel of passive reception

The object of this paper is the process of detecting an aerial object by a radar with active and passive reception channels. The main hypothesis of the study assumed that the introduction of an additional channel of passive reception would increase the conditional probability of correct detection at a fixed value of the conditional probability of false alarms. When detecting an aerial object, it was considered that the signals from aerial object represent a reflected signal after being emitted by the active radar channel elemitted own radio signals. A radar with active and passive reception channels assumes the presence of two channels. The active location channel provides reception of signals reflected from an aerial object, their processing and detection according to the Neumann-Pearson criterion. The passive location channel functions according to the principle of panoramic spectral analysis based on windowed Fourier transforms. The information combining device is intended for the compatible combining of information from active and passive channels of radar reception. At the output of the passive location channel, an output signal is formed, and the coordinates of the aerial object are measured. To ensure the functioning of the radar with active and passive reception channels, it is necessary to ensure the time synchronization of the reception channels. The quality of detection of aerial objects by radar with active and passive signal reception channels was evaluated. The quality indicator defines the conditional probability of correct detection. The dependences of the conditional probability of correct detection are given for a radar with only an active reception channel and a radar with active and passive reception channels. It was established that the introduction of an additional channel of passive reception to the radar increases the conditional probability of correct detection by an average of (20–40) %, depending on the signal-to-noise ratio.

YOLO-MMS for aerial object detection model based on hybrid feature extractor and improved multi-scale prediction

YOLO-MMS for aerial object detection model based on hybrid feature extractor and improved multi-scale prediction

MobVGG: Ensemble technique for birds and drones prediction

Detection of aerial activities, including drones and birds, has practical implications for automating bird surveys and developing radar systems for aerial object collision detection. Convolutional neural networks (CNNs) have been extensively utilized for image recognition and classification tasks, albeit prior research predominantly focuses on single-class 'drone' classification. However, a gap persists in achieving high accuracy for multi-class classification. To address the limitations of traditional CNNs, such as vanishing gradients and the necessity for numerous layers, this study introduces a novel model termed "MobVGG.” This model combines the architectures of MobileNetV2 and VGG16 to accurately classify images as either 'bird' or 'drone'. The dataset comprises 4212 images for each category of 'bird' and 'drone'. The stringent methodology was applied for dataset preparation and model training to ensure the reliability of the findings. Comparative analysis with previous research demonstrates that the proposed MobVGG model, trained on both 'bird' and 'drone' images, achieves superior accuracy (96 %) compared to benchmark studies. Our paper targets researchers and graduate students as its primary audience.

Principles of stereo reconstruction of aerial objects using stationary cameras

ABSTRACT An overview is given here of the principles and mathematics of stereo reconstruction of objects in the sky using stationary cameras with an emphasis on meteorological applications. Through its Atmospheric Radiation Measurement program, the Department of Energy has operated stereo-photogrammetric cameras since 2017 as part of an effort to measure the life-cycle properties of clouds. At the core of that technology is stereo reconstruction, which calculates the real-world position of an object from the location of the object’s image in two cameras’ photographs. Here, stereo reconstruction is stripped down to its basic elements and presented using conventions tailored to applications in atmospheric science. In addition, the resulting equations are used to illustrate the high sensitivity of reconstructed cloud positions to errors in the cameras’ Euler angles. The interested reader will find here a self-contained guide to performing stereo reconstructions using distortion-corrected images from a pair of calibrated, stationary cameras, as well as a demonstration of the need for high accuracy in the measurement of camera properties and orientations.

INVESTIGATION OF DEEP LEARNING MODELS BASED ON SINGLE-LAYER SimpleRNN, LSTM AND GRU NETWORKS FOR RECOGNIZING SOUNDS OF UAV DISTANCES

In recent years, the potential risks posed by easily moving objects have highlighted the need for intelligent surveillance systems in protected areas, primarily to ensure the safety of human lives. Among the most common of these objects are unmanned aerial vehicles (UAVs). Recent advances in deep learning techniques for recognizing audio signals have made these techniques effective in identifying moving or aerial objects, especially those powered by engines. And the growing deployment of UAVs has made their rapid recognition in various suspicious or unauthorized circumstances critical. Detecting suspicious drone flights, especially in restricted areas, remains a significant research challenge. It is vital to perform the task of determining their distance in order to quickly detect drones approaching people in such protected areas. Therefore, this paper aims to study the research question of recognizing UAV audio data from different distances. That is, recognizing drone audio at different distances was experimentally studied using Simple RNN, LSTM and GRU based deep learning models. The main objective of this study is based on finding one of the capable types of recurrent network for the task of recognizing UAV audio data at different distances. During the experimental study, the recognition abilities of Single-layer Simple RNN, LSTM and GRU recurrent network types were studied from two basic directions: with recognition accuracy curves and classification reports. As a result, LSTM and GRU based models showed high recognition ability for these types of audio signals. It was noted that UAVs can reliably predict distances greater than 10 meters based on the proposed deep learning architecture.

YOLOv5_mamba: unmanned aerial vehicle object detection based on bidirectional dense feedback network and adaptive gate feature fusion

Addressing the problem that the object size in Unmanned Aerial Vehicles (UAVs) aerial images is too small and contains limited feature information, leading to existing detection algorithms having less than ideal performance in small object detection, we propose a UAV aerial object detection system named YOLv5_mamba based on bidirectional dense feedback network and adaptive gate feature fusion. This paper improves the You Only Look Once Version 5 (YOLOv5) algorithm by firstly introducing the Faster Implementation of CSP Bottleneck with 2 convolutions (C2f) module from YOLOv8 into the backbone network to enhance the feature extraction capability of the backbone network. Furthermore, the mamba module and C2f module are introduced to construct a bidirectional dense feedback network to enhance the transfer of contextual information in the neck part. Thirdly, an adaptive gate feature fusion network is proposed to improve the head part of YOLOv5 and enhance its final detection capability. Experimental results on the public UAV aerial dataset VisDrone2019 demonstrate that the proposed algorithm improves the detection accuracy by 9.3% compared to the original YOLOv5 baseline network, showing better detection performance for small objects. For the UCAS_AOD dataset, the proposed algorithm outperforms YOLOv5-s by 9%. In the case of the DIOR dataset, the proposed algorithm exceeds YOLOv5-s by 12%.

YOLO V3 AND CCN FOR THE TRACKING AND CLASSIFICATION OF AERIAL OBJECT AND DRONES

The goal of this study is to give headways in aeronautical article ID that will help with making recognitions that are both more exact and more precise. Specifically, we revamp the meaning of the article recognition anchor enclose request to remember turns for expansion to level and width, and besides, we make it conceivable to have erratic four corner point structures. Furthermore, the consideration of new anchor boxes gives the model additional adaptability to address protests that are focused at a pivot of turn that gives a 45-degree point. By accomplishing these results, we can make an organization that considers negligible tradeoffs about speed and unwavering quality, while likewise giving more exact restrictions. The latest ways to deal with PC vision and article acknowledgment are for the most part dependent on brain organizations and different advances that utilize profound learning. This powerful field of study is utilized in various applications, including military and observation, aeronautical photography, independent driving, and airborne perception. To precisely locate the location of an item, contemporary object identification techniques make use of bounding boxes that are drawn over the object and have a rectangular form (horizontal and vertical). These orthogonal bounding boxes do not consider the posture of the object, which leads to a decrease in the amount of object localization and restricts subsequent tasks such as object comprehension and tracking. We have used the DOTA dataset to present all of the results, demonstrating the value of flexible object boundaries, particularly with rotated and non-rectangular objects. We have also achieved an accuracy of 98.47% for the detection and classification of aerial objects, with forty percent of the data being used for training and the remaining twenty percent being used for testing. There was a minimum of 2.8 seconds of processing time required for the whole program to be executed to categorize all of the aerial items that were parked on the base.

NEFELI: A deep-learning detection and tracking pipeline for enhancing autonomy in advanced air mobility

Efficient detection and accurate collision estimation for non-cooperative aerial vehicles are crucial for the realization of fully autonomous aircraft and Advanced Air Mobility (AAM). This paper introduces NEFELI, a machine learning software utilizing optical sensors to detect and track non-cooperative aerial vehicles. NEFELI's detector employs an enhanced YOLOv5-large model, strengthened with a sliced inference step to enhance detection capabilities for distant, diminutive objects. Furthermore, NEFELI introduces several innovations in its tracking component. A key advancement lies in the creation and utilization of the first large-scale re-identification (Re-ID) dataset of aerial objects. This dataset is used to train the deep learning appearance (Re-ID) model of the tracking module and integrates appearance information into the detection and tracking pipeline, resulting in more robust and reliable tracking performance. Moreover, the tracking model combines the deep learning appearance model with a Kalman Filter-based motion model to address the challenge of precisely tracking distant aerial objects. Notably, an extensive comparative analysis that was conducted showed that NEFELI outperforms state-of-the-art detection and tracking models in terms of Higher Order Tracking Accuracy (HOTA) metric, ID switches, and Association Accuracy (AssA) by a wide margin. A crucial aspect of this work is NEFELI's software architecture design, which enables efficient implementation on a low SWaP (Size, Weight, and Power) edge Graphic Processing Unit (GPU). To further showcase NEFELI's generalization capabilities and edge implementation performance, real-world flight experiments with small UAVs were carried out. The experimental results demonstrate NEFELI's ability to detect and track small UAVs at distances of up to 145 m in real-time speed of 6.7 fps.

Small Object Detection in UAV Images Based on YOLOv8n

With the rapid development of unmanned aerial vehicle (UAV) technology, there is an urgent need for high-performance aerial object detection algorithms that are tailored for deployment on drones with limited computing capabilities. This paper proposes a series of improvements to the state-of-the-art YOLOv8 object detector to enhance its detection accuracy and speed for small, partially occluded objects in complex environments. Specifically, we introduce multi-scale feature fusion through additional detection layers, employ conditionally parameterized convolutions to increase representational capacity, and import a dynamic non-monotonic loss function named Wise-IoU to enable more effective regression of bounding boxes. Experiments conducted on the large-scale UAV benchmark dataset VisDrone demonstrate that the improved model achieves state-of-the-art accuracy of 37.6% mAP with 3 M parameters, outperforming other lightweight YOLO detectors and two-stage detectors like Faster R-CNN. The improved model also reaches 40 FPS on an embedded edge device, validating its efficiency and suitability for real-time UAV applications. Through comprehensive quantitative experiments and visual results, this work provides valuable insights and techniques to tailor object detection algorithms for robust and efficient deployment on UAVs with limited onboard computing power.

MAFPN: a mixed local-global attention feature pyramid network for aerial object detection

ABSTRACT Aerial object detection (AOD) aims at locating and categorizing aerial targets in remote sensing image, which has wide practical applications. Mainstream feature pyramid network (FPN) has made great progresses in multi-scale natural objects detection, which is also successfully extended into AOD. However, existing FPNs still have some shortcomings in discriminative feature extraction and fusion for aerial objects due to diverse object scales and dense small objects. To address these limitations, this letter proposes a novel mixed local-global attention FPN (MAFPN). The core components of MAFPN are cross-scale transformer (CST) module and multi-scale spatial fusion (MSSF) module, which are equipped at the feature extraction and fusion of FPN, respectively. Specifically, the CST module is devised to capture global semantic information and cross-scale similarity by exploiting self-attention in multi-scale features. The MSSF module utilizes the spatial attention maps as fusion weights, and adaptively fuses multi-scale features between adjacent two levels. Such mixed attention approach enhances the multi-scale feature learning in dense small objects. Besides, to further balance the influences of different object scales, we propose an area-guided dot-IOU (ADIOU) regression loss, which takes into account the size of objects and enables the model to prioritize the centre position of small objects. Experimental results on public datasets demonstrate the superiority of proposed method.

Recent Advances for Aerial Object Detection: A Survey

Aerial object detection, as object detection in aerial images captured from an overhead perspective, has been widely applied in urban management, industrial inspection, and other aspects. However, the performance of existing aerial object detection algorithms is hindered by variations in object scales and orientations attributed to the aerial perspective. This survey presents a comprehensive review of recent advances in aerial object detection. We start with some basic concepts of aerial object detection and then summarize the five imbalance problems of aerial object detection, including scale imbalance, spatial imbalance, objective imbalance, semantic imbalance, and class imbalance. Moreover, we classify and analyze relevant methods and especially introduce the applications of aerial object detection in practical scenarios. Finally, the performance evaluation is presented on two popular aerial object detection datasets VisDrone-DET and DOTA, and we discuss several future directions that could facilitate the development of aerial object detection.

EL-YOLO: An efficient and lightweight low-altitude aerial objects detector for onboard applications

Existing deep learning-based low-altitude small object detectors are typically complex in model architecture and demand substantial computational resources, making deployment for real-time detection tasks on edge computing devices challenging. To address this issue, we proposed EL-YOLO, an efficient and lightweight onboard applications object detector. Initially, in order to maintain a lightweight design and improve model accuracy, we propose a novel approach called Sparsely Connected Asymptotic Feature Pyramid Network(SCAFPN). This approach aims to eliminate inter-layer interference during feature fusion, thereby enhancing the model’s performance. Furthermore, to establish long-range contextual relationships for small object scale information, we devise a Cross-Space Learning Multi-Head Self-Attention mechanism (CSL-MHSA). To assess EL-YOLO capability for onboard small object detection, we deploy it on the embedded NVIDIA Jetson Xavier Nx platform and employ NVIDIA TensorRT FP16 quantization acceleration. On the VisDrone2019-DET and AI-TOD datasets, EL-YOLO demonstrates a 12.4% and 1.3% improvement in mAP50 compared to YOLOv5s. In comparison with the state-of-the-art YOLOv8s proposed in 2023, it achieves a respective 2.8% and 10.7% increase in mAP50. Moreover, the inference speed for the Small model reaches 24 frames per second, while the Nano model achieves 35 frames per second. This method holds promise for direct integration onto unmanned aerial vehicles for aerial small object detection tasks in the future.

Information Technology for the Operational Processing of Military Content for Commanders of Tactical Army Units

In the modern world, the military sphere occupies a very high place in the life of the country. At the same time, this area needs quick and accurate solutions. This decision can greatly affect the unfolding of events on the battlefield and indicate that they must be used carefully, using all possible means. During the war, the speed and importance of decisions are very important, and we note that the relevance of this topic is growing sharply. The purpose of the work is to create a comprehensive information system that facilitates the work of commanders of tactical units, which organizes the visualization and classification of aerial objects in real-time, the classification of objects for radio-technical intelligence, the structuring of military information and facilitates the perception of military information. The object of research/development is a phenomenon that creates a problematic problem, has the presence of slowing factors in the process of command and control, using teams of tactical links, which can slow down decision-making, as well as affect their correctness. The research/development aims to address emerging bottlenecks in the command-and-control process performed by tactical link teams, providing improved visualization, analysis and work with military data. The result of the work is an information system for processing military data to help commanders of tactical units. This system significantly improves on known officer assistance tools, although it includes a set of programs that have been used in parallel on an as-needed basis. Using modern information technologies and ease of use, the system covers problems that may arise for commanders. Also, each program included in the complex information system has its degree of innovation. The information system for structuring military information is distinguished by the possibility of use on any device. The information system for the visualization and clustering of aerial objects and the information system for the classification of objects for radio technical intelligence are distinguished by their component nature. This means that the application can use sources of input information and provides an API to use other processing information. Regarding the information system for integration into information materials, largely unknown terms and abbreviations are defined, so such solutions, cannot integrate the required data into real documents. Therefore, using this comprehensive information system, the command of tactical units will have the opportunity to improve the quality and achieve the command-and-control process.

Classifying Aerial Objects Based on Risk: A Machine Learning Approach

The spy balloon that flew across the U.S. in February 2023 posed a serious security threat. US security officials have said this balloon tried to gather intelligence by monitoring sensitive military sites, and as a result, the U.S. government began more closely scrutinizing its airspace to better categorize aerial objects and detect threats. However, the airspace is filled with a myriad of aerial objects, making the problem of classifying and risk determination very challenging. We hypothesize that if we label aerial objects based on the risks they pose, then a Machine Learning algorithm can be made to learn and predict the risks of previously unknown or unseen aerial objects. Currently, there are no known single datasets that contain both old and newer aerial objects, such as drones, planes, etc., nor do datasets have labels to identify the risks associated with the objects. The goal of this research is twofold: 1) We create a new comprehensive dataset that contains traditional and newer aerial objects. We label the objects as high, medium, low, or no risk based on the threat markers on the aerial objects. 2) We use the MobileNetV2 CNN classification algorithm to validate the dataset and provide accurate results. Advancements in this space can potentially help intelligence agencies and security analysts quickly assess developing scenarios and provide a reliable risk assessment for observed aerial objects.

LENet: Lightweight and Effective Detector for Aerial Object

Aerial object detection is crucial in various computer vision tasks, including video monitoring, early warning systems, and visual tracking. While current methods can accurately detect normal-sized objects, they face challenges distinguishing small objects from cluttered backgrounds. Developing methods that can be deployed on edge devices to achieve fast, accurate, and energy-efficient performance is also an urgent challenge. This paper proposes a network for aerial object detection by incorporating an attention mechanism to enhance feature extraction and elevate the accuracy of aerial moving object detection. Additionally, we optimize the channel dimensions of the feature extraction framework, resulting in a reduction in model parameters, acceleration of inference speed, and alleviation of the computational burden. Ulteriorly, we optimize the Spatial Pyramid Pooling (SPP) module to enhance detection accuracy and processing speed. Inspired by the ResNet and RepVGG structures, we design a feature fusion module to combine early-extracted features, improving speed and accuracy. Based on the design mentioned above principles, we develop a neural network method with an impressively small model size of only [Formula: see text][Formula: see text]M. The proposed approach achieves state-of-the-art performance on five benchmark datasets. Besides its superior performance, our method demonstrates excellent throughput on edge computing devices. Experimental results show that even when running on low-performance computing devices, the CPU and GPU temperatures remain below 50∘C and achieve a detection speed of 14.8 frames per second (fps) and power consumption of only 2.9[Formula: see text]W. These findings suggest that a high-accuracy, low-power, low-latency, and low-memory footprint aerial object detection solution is achievable.

ОСОБЕННОСТИ ОБРАБОТКИ БОЛЬШИХ ОБЪЕМОВ ДАННЫХ ПРИ УПРАВЛЕНИИ БЕСПИЛОТНЫМИ ЛЕТАТЕЛЬНЫМИ АППАРАТАМИ НА ОСНОВЕ ИМИТАЦИОННОГО МОДЕЛИРОВАНИЯ

The paper discusses the principles of constructing a simulation model to study the influence of the type of transactions to a database management system on the reception and assimilation of data from unmanned aerial objects. The proposed study examines the interaction of unmanned aerial objects with a data receiver when implementing the monitoring process. The interaction time of the unmanned aerial vehicle with the dispatch center is modeled based on various types of transactions to the database, taking into account the differentiation of the database management system server load and their impact on the feedback time from the unmanned aerial vehicle, which will allow assessing the risks of control in an online mode.

AFDet: alignment and focusing for aerial object detection

ABSTRACT Object detection in aerial images has become a focus in recent years due to the expansion of its application fields. Aerial objects have the characteristics of multi-scale, arbitrary angle and dense arrangement, which brings considerable challenges to the task. Based on the efficient one-stage detectors, most methods design modules to adaptively align the feature receptive field with the anchors to improve the accuracy of the bounding box regression and calculate loss according to parameter bias. Current methods are not adequately data-driven in feature optimization and the angle parameter regression faces boundary problems. To handle these problems, in this letter, we propose an Align-Focus detector (AFDet) for aerial image object detection. We introduce the ideology of deformable multi-head self-attention to feature optimization and design a new Align-Focus Module (AFM) which can sensitively focus feature encoding response to the real texture area of the objects. In addition, we apply KF-IOU Loss to solve the boundary problem. We conduct necessary experiments on DOTA and HRSC2016 datasets, and AFDet achieves the highest mAP performance compared to the existing methods.

Zero-Shot Aerial Object Detection with Visual Description Regularization

Existing object detection models are mainly trained on large-scale labeled datasets. However, annotating data for novel aerial object classes is expensive since it is time-consuming and may require expert knowledge. Thus, it is desirable to study label-efficient object detection methods on aerial images. In this work, we propose a zero-shot method for aerial object detection named visual Description Regularization, or DescReg. Concretely, we identify the weak semantic-visual correlation of the aerial objects and aim to address the challenge with prior descriptions of their visual appearance. Instead of directly encoding the descriptions into class embedding space which suffers from the representation gap problem, we propose to infuse the prior inter-class visual similarity conveyed in the descriptions into the embedding learning. The infusion process is accomplished with a newly designed similarity-aware triplet loss which incorporates structured regularization on the representation space. We conduct extensive experiments with three challenging aerial object detection datasets, including DIOR, xView, and DOTA. The results demonstrate that DescReg significantly outperforms the state-of-the-art ZSD methods with complex projection designs and generative frameworks, e.g., DescReg outperforms best reported ZSD method on DIOR by 4.5 mAP on unseen classes and 8.1 in HM. We further show the generalizability of DescReg by integrating it into generative ZSD methods as well as varying the detection architecture. Codes will be released at https://github.com/zq-zang/DescReg.

An Integrated Detection and Multi-Object Tracking Pipeline for Satellite Video Analysis of Maritime and Aerial Objects

Optical remote sensing videos, as a new source of remote sensing data that has emerged in recent years, have significant potential in remote sensing applications, especially national defense. In this paper, a tracking pipeline named TDNet (tracking while detecting based on a neural network) is proposed for optical remote sensing videos based on a correlation filter and deep neural networks. The pipeline is used to simultaneously track ships and planes in videos. There are many target tracking methods for general video data, but they suffer some difficulties in remote sensing videos with low resolution and those influenced by weather conditions. The tracked targets are usually misty. Therefore, in TDNet, we propose a new multi-target tracking method called MT-KCF and a detecting-assisted tracking (i.e., DAT) module to improve tracking accuracy and precision. Meanwhile, we also design a new target recognition (i.e., NTR) module to recognise newly emerged targets. In order to verify the performance of TDNet, we compare our method with several state-of-the-art tracking methods on optical video remote sensing data sets acquired from the Jilin No. 1 satellite. The experimental results demonstrate the effectiveness and the state-of-the-art performance of the proposed method. The proposed method can achieve more than 90% performance in terms of precision for single-target tracking tasks and more than 85% performance in terms of MOTA for multi-object tracking tasks.