{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,7]],"date-time":"2026-07-07T04:31:15Z","timestamp":1783398675547,"version":"3.54.6"},"reference-count":60,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T00:00:00Z","timestamp":1753833600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurorobot."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>Accurate vehicle analysis from aerial imagery has become increasingly vital for emerging technologies and public service applications such as intelligent traffic management, urban planning, autonomous navigation, and military surveillance. However, analyzing UAV-captured video poses several inherent challenges, such as the small size of target vehicles, occlusions, cluttered urban backgrounds, motion blur, and fluctuating lighting conditions which hinder the accuracy and consistency of conventional perception systems. To address these complexities, our research proposes a fully end-to-end deep learning\u2013driven perception pipeline specifically optimized for UAV-based traffic monitoring. The proposed framwork integrates multiple advanced modules: RetinexNet for preprocessing, segmentation using HRNet to preserve high-resolution semantic information, and vehicle detection using the YOLOv11 framework. Deep SORT is employed for efficient vehicle tracking, while CSRNet facilitates high-density vehicle counting. LSTM networks are integrated to predict vehicle trajectories based on temporal patterns, and a combination of DenseNet and SuperPoint is utilized for robust feature extraction. Finally, classification is performed using Vision Transformers (ViTs), leveraging attention mechanisms to ensure accurate recognition across diverse categories. The modular yet unified architecture is designed to handle spatiotemporal dynamics, making it suitable for real-time deployment in diverse UAV platforms.<\/jats:p><\/jats:sec><jats:sec><jats:title>Method<\/jats:title><jats:p>The framework suggests using today\u2019s best neural networks that are made to solve different problems in aerial vehicle analysis. RetinexNet is used in preprocessing to make the lighting of each input frame consistent. Using HRNet for semantic segmentation allows for accurate splitting between vehicles and their surroundings. YOLOv11 provides high precision and quick vehicle detection and Deep SORT allows reliable tracking without losing track of individual cars. CSRNet are used for vehicle counting that is unaffected by obstacles or traffic jams. LSTM models capture how a car moves in time to forecast future positions. Combining DenseNet and SuperPoint embeddings that were improved with an AutoEncoder is done during feature extraction. In the end, using an attention function, Vision Transformer-based models classify vehicles seen from above. Every part of the system is developed and included to give the improved performance when the UAV is being used in real life.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Our proposed framework significantly improves the accuracy, reliability, and efficiency of vehicle analysis from UAV imagery. Our pipeline was rigorously evaluated on two famous datasets, AU-AIR and Roundabout. On the AU-AIR dataset, the system achieved a detection accuracy of 97.8%, a tracking accuracy of 96.5%, and a classification accuracy of 98.4%. Similarly, on the Roundabout dataset, it reached 96.9% detection accuracy, 94.4% tracking accuracy, and 97.7% classification accuracy. These results surpass previous benchmarks, demonstrating the system\u2019s robust performance across diverse aerial traffic scenarios. The integration of advanced models, YOLOv11 for detection, HRNet for segmentation, Deep SORT for tracking, CSRNet for counting, LSTM for trajectory prediction, and Vision Transformers for classification enables the framework to maintain high accuracy even under challenging conditions like occlusion, variable lighting, and scale variations.<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion<\/jats:title><jats:p>The outcomes show that the chosen deep learning system is powerful enough to deal with the challenges of aerial vehicle analysis and gives reliable and precise results in all the aforementioned tasks. Combining several advanced models ensures that the system works smoothly even when dealing with problems like people being covered up and varying sizes.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fnbot.2025.1643011","type":"journal-article","created":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T05:30:37Z","timestamp":1753853437000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":22,"title":["Integrated neural network framework for multi-object detection and recognition using UAV imagery"],"prefix":"10.3389","volume":"19","author":[{"given":"Mohammed","family":"Alshehri","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tingting","family":"Xue","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ghulam","family":"Mujtaba","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yahya","family":"AlQahtani","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nouf Abdullah","family":"Almujally","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ahmad","family":"Jalal","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hui","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,7,30]]},"reference":[{"key":"ref1","author":"Abrar","year":"2019"},{"key":"ref2","first-page":"1","article-title":"Scene semantic recognition based on modified fuzzy c-mean and maximum entropy using object-to-object relations","volume":"9","author":"Ahmad","year":"2021","journal-title":"In IEEE Access"},{"key":"ref001","volume-title":"Region and decision tree-based segmentations for multi-object detection and classification in outdoor scenes.","author":"Ahmed","year":"2024"},{"key":"ref3","first-page":"353","author":"Altch\u00e9","year":"2017"},{"key":"ref002","doi-asserted-by":"publisher","first-page":"80973","DOI":"10.1109\/ACCESS.2023.3299488","article-title":"A Smart Traffic Control System Based on Pixel-Labeling and SORT Tracker","volume":"11","author":"Alonazi","year":"2023","journal-title":"IEEE Access"},{"key":"ref4","volume-title":"Automated body parts estimation and detection using salient maps and Gaussian matrix model","author":"Ayesha","year":"2021"},{"key":"ref5","doi-asserted-by":"crossref","DOI":"10.1109\/ComTech52583.2021.9616681","volume-title":"Smartphone inertial sensors for human locomotion activity recognition based on template matching and codebook generation","author":"Azmat","year":"2021"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"4832","DOI":"10.3390\/s23104832","article-title":"Deep learning techniques for vehicle detection and classification from images\/videos: A survey","volume":"23","author":"Berwo","year":"2023","journal-title":"In Sensors"},{"key":"ref7","volume-title":"Object detection and segmentation for scene understanding via multi-features and random forest","author":"Bisma","year":"2023"},{"key":"ref8","volume-title":"R-CNN based vehicle object detection via segmentation capabilities in road scenes","author":"Bisma","year":"2025"},{"key":"ref9","doi-asserted-by":"crossref","first-page":"6047","DOI":"10.1109\/TNNLS.2021.3080276","article-title":"Vehicle detection from UAV imagery with deep learning: A review","volume":"33","author":"Bouguettaya","year":"2021","journal-title":"In IEEE Trans. Neural Networks Learn. Syst."},{"key":"ref10","first-page":"8504","author":"Bozcan","year":"2020"},{"key":"ref11","first-page":"1","author":"Chaman","year":"2025"},{"key":"ref12","article-title":"Object detection and segmentation for scene understanding via random forest","author":"Chughtai","year":"","journal-title":"ICACS"},{"key":"ref13","doi-asserted-by":"crossref","DOI":"10.1109\/ICACS55311.2023.10089658","volume-title":"Object detection and segmentation for scene understanding via random forest","author":"Chughtai","year":""},{"key":"ref14","article-title":"Traffic surveillance system: robust multiclass vehicle detection and classification","author":"Chughtai","year":"2024","journal-title":"ICACS"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.3390\/wevj15080344","article-title":"Vehicle classification algorithm based on improved vision transformer","volume":"15","author":"Dong","year":"2024","journal-title":"In World Electric Vehicle J."},{"key":"ref16","volume-title":"Robust vehicle detection and tracking model via deep SORT over aerial images","author":"Ghulam","year":"2024"},{"key":"ref17","article-title":"Remote sensing-based traffic monitoring via semantic segmentation and deep learning","author":"Ghulam","year":"2024","journal-title":"INMIC"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"201","DOI":"10.1016\/j.isprsjprs.2022.05.004","article-title":"Density map-based vehicle counting in remote sensing images with limited resolution","volume":"189","author":"Guo","year":"2022","journal-title":"ISPRS J. Photogram. Remote Sensing"},{"key":"ref19","doi-asserted-by":"crossref","DOI":"10.1109\/ComTech52583.2021.9616668","volume-title":"Multi-fusion sensors for action recognition based on discriminative motion cues and random forest","author":"Hafeez","year":"2021"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"116916","DOI":"10.1016\/j.image.2022.116916","article-title":"Advanced retinexnet: A fully convolutional network for low-light image enhancement","volume":"112","author":"Hai","year":"2023","journal-title":"Signal Process."},{"key":"ref21","first-page":"2941","author":"Han","year":"2018"},{"key":"ref22","article-title":"Smart traffic monitoring through drone images via YOLOv5 and Kalman filter","author":"Hanzla","year":"","journal-title":"ICACS"},{"key":"ref23","doi-asserted-by":"crossref","DOI":"10.1109\/ICACS64902.2025.10937840","volume-title":"Intelligent transportation surveillance via YOLOv9 and NASNet over aerial imagery","author":"Hanzla","year":"2025"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"14032","DOI":"10.1038\/s41598-025-96314-x","article-title":"Research on object detection and recognition in remote sensing images based on YOLOv11","volume":"15","author":"He","year":"2025","journal-title":"Sci. Rep."},{"key":"ref28","first-page":"1","author":"Hossain","year":"2019"},{"key":"ref29","first-page":"1","author":"Hou","year":"2019"},{"key":"ref30","article-title":"Densenet: Implementing efficient convnet descriptor pyramids","author":"Iandola","year":"2014","journal-title":"arXiv preprint arXiv"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"119","DOI":"10.32604\/iasc.2022.024812","article-title":"Deep neural network based vehicle detection and classification of aerial images","volume":"34","author":"Kumar","year":"2022","journal-title":"Intell. Automat. Soft Computing"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"212209","DOI":"10.1109\/ACCESS.2020.3040290","article-title":"VAID: an aerial image dataset for vehicle detection and classification","volume":"8","author":"Lin","year":"2020","journal-title":"IEEE Access"},{"key":"ref33","author":"Mahammed","year":"2023"},{"key":"ref34","volume-title":"Artificial neural network for human object interaction system over aerial images","author":"Mahwish","year":"2023"},{"key":"ref35","volume-title":"Hybrid algorithm for multi people counting and tracking for smart surveillance","author":"Mahwish","year":"2021"},{"key":"ref36","article-title":"Unmanned aerial vehicle based multi-person detection via deep neural network models","author":"Mohammed","year":"2025","journal-title":"Front. Neurorobot."},{"key":"ref37","article-title":"UAV-based road traffic monitoring via FCN segmentation and Deepsort for smart cities","author":"Mujtaba","year":"2024","journal-title":"ICIC"},{"key":"ref38","doi-asserted-by":"crossref","DOI":"10.1109\/ICACS64902.2025.10937829","volume-title":"Drone surveillance for intelligent multi-vehicles monitoring and classification","author":"Mujtaba","year":"2025"},{"key":"ref39","doi-asserted-by":"crossref","DOI":"10.1109\/ICACS55311.2023.10089778","volume-title":"Automate appliances via gestures recognition for elderly living assistance","author":"Muneeb","year":"2023"},{"key":"ref40","volume-title":"Integrating semantic segmentation and object detection for multi-object labeling in aerial images","author":"Naseer","year":"2024"},{"key":"ref41","article-title":"Multimodal objects categorization by fusing GMM and multi-layer perceptron","author":"Naseer","year":"2024","journal-title":"ICACS"},{"key":"ref42","doi-asserted-by":"crossref","DOI":"10.1109\/ICACS64902.2025.10937817","volume-title":"Multimodal deep learning framework for enhanced semantic scene classification using RGB-D images","author":"Naseer","year":""},{"key":"ref43","article-title":"Hybrid deep learning aerial framework for road scene objects segmentation and classification","author":"Naseer","year":"","journal-title":"ICACS"},{"key":"ref44","article-title":"Efficient aerial images algorithms over multi-objects labeling and semantic segmentation","author":"Naseer","year":"2024","journal-title":"ICACS"},{"key":"ref45","volume-title":"Tracking and analysis of pedestrian's behavior in public places","author":"Pervaiz","year":"2023"},{"key":"ref46","first-page":"1215","author":"Petrakis","year":"2023"},{"key":"ref47","doi-asserted-by":"publisher","first-page":"47","DOI":"10.3390\/data7040047","article-title":"Dataset: roundabout aerial images for vehicle detection","volume":"7","author":"Puertas","year":"2022","journal-title":"In Data"},{"key":"ref48","first-page":"3275","volume-title":"Multiple pedestrian detection and tracking in night vision surveillance systems","author":"Raza","year":"2023"},{"key":"ref49","volume-title":"Vehicle detection and tracking from aerial imagery via YOLO and centroid tracking","author":"Shuja","year":"2023"},{"key":"ref50","doi-asserted-by":"publisher","first-page":"406","DOI":"10.3390\/drones6120406","article-title":"FRCNN-based reinforcement learning for real-time vehicle detection, tracking and geolocation from UAS","volume":"6","author":"Singh","year":"2022","journal-title":"Drones"},{"key":"ref51","first-page":"5853","volume-title":"Exploiting human pose and scene information for interaction detection","author":"Waheed","year":"2023"},{"key":"ref52","doi-asserted-by":"publisher","first-page":"446","DOI":"10.3390\/rs9050446","article-title":"Gated convolutional neural network for semantic segmentation in high-resolution images","volume":"9","author":"Wang","year":"2017","journal-title":"Remote Sensing"},{"key":"ref53","volume-title":"RGB-D scene classification: A unified framework with vision transformers and contextual models","author":"Waqas","year":"2024"},{"key":"ref54","volume-title":"Indoor scene classification using RGB-D data: A vision transformer and conditional random field approach","author":"Waqas","year":"2024"},{"key":"ref55","volume-title":"Robust object recognition with genetic algorithm and composite saliency map","author":"Waqas","year":""},{"key":"ref56","volume-title":"Dynamic adaptive Gaussian mixture model for multi-object detection over natural scenes","author":"Waqas","year":""},{"key":"ref57","article-title":"Multi-vehicles classification on aerial images using ResNet and transformer networks","author":"Waqas","year":"2025","journal-title":"ICACS."},{"key":"ref58","first-page":"9372","volume-title":"A novel remote sensing recognition using modified GMM segmentation and DenseNet","author":"Waqas","year":""},{"key":"ref59","volume-title":"Perception of natural scenes: Objects detection and segmentations using saliency map with AlexNet","author":"Waqas","year":""},{"key":"ref60","doi-asserted-by":"publisher","first-page":"105241","DOI":"10.1016\/j.cmpb.2019.105241","article-title":"Skin lesion segmentation using high-resolution convolutional neural network","volume":"186","author":"Xie","year":"2020","journal-title":"Comput. Methods Prog. Biomed."},{"key":"ref61","doi-asserted-by":"publisher","first-page":"1141","DOI":"10.1007\/s11263-019-01266-1","article-title":"The unmanned aerial vehicle benchmark: object detection, tracking and baseline","volume":"128","author":"Yu","year":"2020","journal-title":"Int. J. Computer Vision"}],"container-title":["Frontiers in Neurorobotics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2025.1643011\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T05:30:45Z","timestamp":1753853445000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2025.1643011\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,30]]},"references-count":60,"alternative-id":["10.3389\/fnbot.2025.1643011"],"URL":"https:\/\/doi.org\/10.3389\/fnbot.2025.1643011","relation":{},"ISSN":["1662-5218"],"issn-type":[{"value":"1662-5218","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,30]]},"article-number":"1643011"}}