UAVs and Neural Networks for search and rescue missions

TL;DR

This paper introduces a neural network-based method for detecting objects like humans, cars, and fire in UAV aerial images, utilizing dataset creation, assisted labeling, and data augmentation to improve search and rescue operations.

Contribution

It presents a novel integrated pipeline combining classic image processing with neural networks for object detection and dataset augmentation in UAV imagery.

Findings

Effective object detection in UAV images demonstrated

Enhanced dataset with automatic labeling improves model training

Performance evaluation of various neural networks conducted

Abstract

In this paper, we present a method for detecting objects of interest, including cars, humans, and fire, in aerial images captured by unmanned aerial vehicles (UAVs) usually during vegetation fires. To achieve this, we use artificial neural networks and create a dataset for supervised learning. We accomplish the assisted labeling of the dataset through the implementation of an object detection pipeline that combines classic image processing techniques with pretrained neural networks. In addition, we develop a data augmentation pipeline to augment the dataset with automatically labeled images. Finally, we evaluate the performance of different neural networks.