A Comprehensive Review on Computer Vision Analysis of Aerial Data

TL;DR

This comprehensive review covers key computer vision tasks in aerial data analysis, comparing architectures, datasets, and challenges, and discusses future research directions in the field.

Contribution

It provides an extensive overview of methodologies, datasets, and challenges in aerial data computer vision, highlighting gaps and future research opportunities.

Findings

Comparison of hyperparameters across architectures

Discussion of datasets and evaluation metrics

Identification of unresolved challenges and future directions

Abstract

With the emergence of new technologies in the field of airborne platforms and imaging sensors, aerial data analysis is becoming very popular, capitalizing on its advantages over land data. This paper presents a comprehensive review of the computer vision tasks within the domain of aerial data analysis. While addressing fundamental aspects such as object detection and tracking, the primary focus is on pivotal tasks like change detection, object segmentation, and scene-level analysis. The paper provides the comparison of various hyper parameters employed across diverse architectures and tasks. A substantial section is dedicated to an in-depth discussion on libraries, their categorization, and their relevance to different domain expertise. The paper encompasses aerial datasets, the architectural nuances adopted, and the evaluation metrics associated with all the tasks in aerial data analysis. Applications of computer vision tasks in aerial data across different domains are explored, with case studies providing further insights. The paper thoroughly examines the challenges inherent in aerial data analysis, offering practical solutions. Additionally, unresolved issues of significance are identified, paving the way for future research directions in the field of aerial data analysis.