Table 2 Analysis of related work for disaster image classification models and algorithms

[17, 23]

Extracting geo-located images from Tweets to support emergency response

Developed a system called IMEXT (Image Extraction from tweets) to extract the geo-located images from tweets for supporting the emergency response

Tweets posted the two days after the earthquake that occurred in central Italy in August 2016

[7]

Study of how people use social media in disasters situations

Hurricane Sandy images by Instagram and Twitter users

[23, 24]

Develop a system (Image4Act) that provides end-to-end social media image processing

The system collects, denoising, and classifying imagery content posted on social media platforms to help humanitarian organizations gain situational awareness and launch relief operations

Cyclone Debbie hit Queensland, Australia in March 2017

[8]

The images on the ground and posted on social media can offer more reliable and valuable information for improving situational awareness

Hurricane Sandy

[9]

Build a machine learning model to classifying fire and non-fire images

California Rim Fire

[10]

Employ machine learning techniques to analyze images posted on social media platforms during natural disasters to determine the level of damage caused by disasters

Typhoon

Ruby typhoon/Hagupit,

Nepal Earthquake, Ecuador Earthquake,

Hurricane Matthew

[25]

Presented a large amount of multimodal Twitter dataset related to natural disasters

Seven disaster events in 2017

[26]

Designed and implemented flood images retrieval pipeline that using deep neural network for comparing various features for image retrieval

No case studies