Disaster Assessment from Satellite Imagery by Analysing Topographical Features Using Deep Learning

Dotel, Saramsha; Shrestha, Avishekh; Bhusal, Anish; Pathak, Ramesh; Shakya, Aman; Panday, Sanjeeb Prasad

doi:10.1145/3388818.3389160

Cited by 22 publications

(12 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A semantic segmentation NN known as U -Net along with the ResNet model was used to determine the damage caused to roads for the Digital Globe satellite images of Hurricane Harvey. An accuracy of 0.845 and F1 -score 0f 0.675 was obtained [10]. VGG-16 CNN/ Multilayer perceptron was used to classify time period and urgency for Hurricane Harvey 2017 images [11].…”

Section: Related Workmentioning

confidence: 99%

“…It becomes impossible to prevent an extreme weather event when the speed of the wind increases the above threshold. Satellite images are gaining popularity for monitoring of hurricanes [3]. Satellite images help in assessing the situation by providing an aerial view.…”

Section: Introductionmentioning

confidence: 99%

“…Li et al[15] obtained an accuracy of 88.3 % while nding out the damage caused due to Hurricane Sandy. Dotel et al[3] obtained an accuracy of 84.5 % for damage detection of Hurricane Harvey. Doshi et al[9] obtained a F1 score of 81.2% for Hurricane Harvey damage detection.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Convolutional neural network based hurricane damage detection using satellite images

et al. 2022

View full text Add to dashboard Cite

Huge swirling storms known as hurricanes are tropical storms appearing in the North Atlantic Ocean and Northeast Paci c that result in winds of 120 km/hour and higher. The winds occurring during hurricanes are catastrophic resulting in immense damage to human life and property. Rapid assessment of damage caused by hurricanes is extremely important for the rst responders. But this process is usually slow, expensive, labor intensive and prone to errors. The advancements in remote sensing and computer vision help in observing Earth at a different scale. In this paper, a Convolutional Neural Network model has been designed that assesses the damage caused to buildings of post hurricane satellite images. The images have been classi ed as Damaged and Undamaged. The model is composed of ve convolutional layers, ve pooling layers, one attening layer, one dropout layer and two dense layers. Hurricane Harvey dataset consisting of 23000 images of size 128 X 128 pixels has been used in this paper. The proposed model performed best at learning rate of 0.00001 and 30 epochs with the Adam optimizer obtaining an accuracy of 0.95, precision of 0.97, recall of 0.96 and F1-score of 0.96. It also achieved the best accuracy and minimum loss.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Convolutional neural network based hurricane damage detection using satellite images

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Saramsha Dotel et al (2020) proposed a deep learning-based landscape monitoring strategy for disaster management. The regions and landscapes affected and changes have been emphasized by the monitoring strategy.…”

Section: Literature Reviewmentioning

confidence: 99%

Advanced deep learning method for Aerial image segmentation of landscape changes in pre-and post-disaster scenarios

2022

Global NEST: The International Journal

View full text Add to dashboard Cite

<p>The precise analysis of conditions in the landscape before and aftermath of the disaster is a mandatory challenge in aerial image landscape monitoring. The change in patterns of landscape, damaged pathways, and damaged areas will have a major impact without monitoring and redevelopment. Therefore, semantic segmentation of the landscape is required in order to analyze the changes and avoid other risks in pre-and post-disaster scenarios. To address these queries a deep learning-based landscape monitoring method is presented in this work. A Gated Shaped Convolution Neural Network is utilized for the semantic segmentation of the aerial landscape images. Initially the aerial image undergoes pre-processing with the process of dilation and GSCNN emphasize the shaped and boundary masks of the affected landscape. To choose the best possible pathways and solutions for development the GSCNN undergoes the particle swarm optimization. In the present study, the proposed PSO-GSCNN is evaluated by comparing the accuracy, precision, and recall of the proposed method with Restricted Boltzmann Machines (RBMs), Convolutional Neural Networks, and Fuse-Net segmentation methods. In comparison with conventional RSM, CNN, and Fuse-Net, the accuracy rate of the model is 97.65%, the precision rate is 98.21%, and the recall rate is 97.23%. This technique has achieved 97.66% accuracy, 97% precision, and 96% recall, all higher than the existing methods.</p>

show abstract

mentioning

confidence: 99%

Convolutional Neural Network based Hurricane Damage Detection using Satellite Images

Kaur

Gupta

Singh

et al. 2021

Preprint

View full text Add to dashboard Cite

Huge swirling storms known as hurricanes are tropical storms appearing in the North Atlantic Ocean and Northeast Pacific that result in winds of 120 km/hour and higher. The winds occurring during hurricanes are catastrophic resulting in immense damage to human life and property. Rapid assessment of damage caused by hurricanes is extremely important for the first responders. But this process is usually slow, expensive, labor intensive and prone to errors. The advancements in remote sensing and computer vision help in observing Earth at a different scale. In this paper, a Convolutional Neural Network model has been designed that assesses the damage caused to buildings of post hurricane satellite images. The images have been classified as Damaged and Undamaged. The model is composed of five convolutional layers, five pooling layers, one flattening layer, one dropout layer and two dense layers. Hurricane Harvey dataset consisting of 23000 images of size 128 X 128 pixels has been used in this paper. The proposed model performed best at learning rate of 0.00001 and 30 epochs with the Adam optimizer obtaining an accuracy of 0.95, precision of 0.97, recall of 0.96 and F1-score of 0.96. It also achieved the best accuracy and minimum loss.

show abstract

Disaster Assessment from Satellite Imagery by Analysing Topographical Features Using Deep Learning

Cited by 22 publications

References 7 publications

Convolutional neural network based hurricane damage detection using satellite images

Convolutional neural network based hurricane damage detection using satellite images

Advanced deep learning method for Aerial image segmentation of landscape changes in pre-and post-disaster scenarios

Convolutional Neural Network based Hurricane Damage Detection using Satellite Images

Contact Info

Product

Resources

About