Surveying Multi-object Classification Methods for UAV Imagery Analysis

Purpose: This paper presents a comprehensive review of multi-object classification methods for Unmanned Aerial Vehicle (UAV) imagery by examining the evolution of traditional machine learning and modern deep learning techniques. The objectives are to determine existing approaches, compare their performance, identify research gaps, and develop a guide for accurate and efficient UAV image analysis. Design/Methodology/Approach: The research offered a comprehensive, cutting-edge object categorisation system for aerial images that incorporates single-stage, two-stage, and advanced-stage detectors. Many recent state-of-the-art object classification algorithms are designed to classify one object at a time. This necessitates new processing and analysis techniques that can fully utilise the enormous volume of gathered information. Research Limitation: The review is limited to published literature and benchmark datasets available during the study period. Variations in datasets, evaluation protocols, experimental settings, and reported performance metrics limit direct comparison across methods. Findings: The review found that deep learning-based methods significantly outperform traditional machine learning techniques in detection accuracy, robustness, and feature-learning capabilities. CNN-based architectures dominate current UAV object classification research, while attention mechanisms and multimodal approaches further improve performance in challenging environments. The analysis also identified persistent challenges, including object occlusion, scale variation, complex backgrounds, limited annotated datasets, computational complexity, and inconsistent reporting of evaluation metrics. Practical Implication: The findings provide valuable guidance for researchers, practitioners, and developers in selecting suitable object classification techniques for UAV applications. Social Implication: It has an impact on precision agriculture, environmental monitoring, disaster management, traffic surveillance, urban planning, and security Originality/Value: Unlike existing surveys, it covers everything from basic traditional algorithms to advanced deep learning algorithms, methodologies, frameworks, new hypotheses, procedures, and applications, although earlier studies focused on single-object detection, particularly a specific target.

Authors

Publication Details

Journal
African Journal Of Applied Research
Published
2026-11-18
DOI
https://doi.org/10.26437/ngfzsa84
Primary Topic
Advanced Neural Network Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Surveying Multi-object Classification Methods for UAV Imagery Analysis

A. A. Khurshid, S. V. Kuntawar
3 links
African Journal Of Applied Research
Advanced Neural Network Applications
article

Surveying Multi-object Classification Methods for UAV Imagery Analysis

A. A. Khurshid, S. V. Kuntawar
article en

Abstract

Purpose: This paper presents a comprehensive review of multi-object classification methods for Unmanned Aerial Vehicle (UAV) imagery by examining the evolution of traditional machine learning and modern deep learning techniques. The objectives are to determine existing approaches, compare their performance, identify research gaps, and develop a guide for accurate and efficient UAV image analysis. Design/Methodology/Approach: The research offered a comprehensive, cutting-edge object categorisation system for aerial images that incorporates single-stage, two-stage, and advanced-stage detectors. Many recent state-of-the-art object classification algorithms are designed to classify one object at a time. This necessitates new processing and analysis techniques that can fully utilise the enormous volume of gathered information. Research Limitation: The review is limited to published literature and benchmark datasets available during the study period. Variations in datasets, evaluation protocols, experimental settings, and reported performance metrics limit direct comparison across methods. Findings: The review found that deep learning-based methods significantly outperform traditional machine learning techniques in detection accuracy, robustness, and feature-learning capabilities. CNN-based architectures dominate current UAV object classification research, while attention mechanisms and multimodal approaches further improve performance in challenging environments. The analysis also identified persistent challenges, including object occlusion, scale variation, complex backgrounds, limited annotated datasets, computational complexity, and inconsistent reporting of evaluation metrics. Practical Implication: The findings provide valuable guidance for researchers, practitioners, and developers in selecting suitable object classification techniques for UAV applications. Social Implication: It has an impact on precision agriculture, environmental monitoring, disaster management, traffic surveillance, urban planning, and security Originality/Value: Unlike existing surveys, it covers everything from basic traditional algorithms to advanced deep learning algorithms, methodologies, frameworks, new hypotheses, procedures, and applications, although earlier studies focused on single-object detection, particularly a specific target.

African Journal Of Applied ResearchVol. 12(8)
Advanced Neural Network Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon