Robust human detection system in flood related images with data augmentation

Floods are one of the major natural disasters that are very common in monsoon countries like India. They cause enormous damage, both to properties and to human lives. It is very essential to detect humans in flooded environments, which plays an important role in rescue management, disaster managemen...

Full description

Saved in:
Bibliographic Details
Published in:Multimedia tools and applications 2023-03, Vol.82 (7), p.10661-10679
Main Authors: M, Dhanushree, S, Chitrakala, C.M., Bhatt
Format: Article
Language:English
Subjects:
Algorithms
Computer Communication Networks
Computer Science
Data augmentation
Data Structures and Information Theory
Datasets
Disaster management
Flood management
Floods
Multimedia Information Systems
Natural disasters
Object recognition
Robustness
Special Purpose and Application-Based Systems
Weather
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Floods are one of the major natural disasters that are very common in monsoon countries like India. They cause enormous damage, both to properties and to human lives. It is very essential to detect humans in flooded environments, which plays an important role in rescue management, disaster management, flood assessment, etc. Detection in a flooded environment is particularly challenging due to the weather conditions, which affect the efficiency of the detection system. The limited availability of proper datasets with natural weather changes also affects the robustness of human object detection in flooded images. In this paper, data augmentation techniques are introduced in order to mimic the changing weather conditions such as rainy, cloudy, and foggy days during flood times, and a HOG-based Robust Human Object Detection (HOG_based_RHOD) algorithm is proposed and has been demonstrated on the augmented dataset. The proposed HOG_based_RHOD algorithm detects human objects in flood-related images, demonstrating its robustness in a variety of challenging weather conditions.
ISSN:1380-7501
1573-7721
DOI:10.1007/s11042-022-13760-9