Corpus ID: 226964894

Street to Cloud: Improving Flood Maps With Crowdsourcing and Semantic Segmentation

  title={Street to Cloud: Improving Flood Maps With Crowdsourcing and Semantic Segmentation},
  author={Veda Sunkara and Matthew Purri and B. L. Saux and Jennifer Adams},
To address the mounting destruction caused by floods in climate-vulnerable regions, we propose Street to Cloud, a machine learning pipeline for incorporating crowdsourced ground truth data into the segmentation of satellite imagery of floods. We propose this approach as a solution to the labor-intensive task of generating high-quality, hand-labeled training data, and demonstrate successes and failures of different plausible crowdsourcing approaches in our model. Street to Cloud leverages… Expand
1 Citations

Figures and Tables from this paper

Enhancement of Detecting Permanent Water and Temporary Water in Flood Disasters by Fusing Sentinel-1 and Sentinel-2 Imagery Using Deep Learning Algorithms: Demonstration of Sen1Floods11 Benchmark Datasets
  • Yanbing Bai, Wenqi Wu, +6 authors S. Koshimura
  • Computer Science
  • Remote. Sens.
  • 2021
Identifying permanent water and temporary water in flood disasters efficiently has mainly relied on change detection method from multi-temporal remote sensing imageries, but estimating the water typeExpand


Integration of Crowdsourced Images, USGS Networks, Remote Sensing, and a Model to Assess Flood Depth during Hurricane Florence
This study integrates flood depth derived from crowdsourced images with U.S. Geological Survey ground-based observation networks, a remote sensing product, and a model during Hurricane Florence to assess flood depth in areas impacted by Hurricane Florence. Expand
Seeing Through the Clouds With DeepWaterMap
The next-generation surface water mapping model, DeepWaterMapV2, is presented, which uses improved model architecture, data set, and a training setup to create surface water maps at lower cost, with higher precision and recall, and is memory efficient for large inputs. Expand
Sen1Floods11: a georeferenced dataset to train and test deep learning flood algorithms for Sentinel-1
The results suggest deep learning models for flood detection of radar data can outperform threshold based remote sensing algorithms, and perform better with training labels that include flood water specifically, not just permanent surface water. Expand
Bridging the Domain Gap for Ground-to-Aerial Image Matching
  • Krishna Regmi, M. Shah
  • Computer Science
  • 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
  • 2019
This work synthesizes an aerial representation of a ground-level panorama query and uses it to minimize the domain gap between the two views, and fuse the complementary features from a synthesized aerial image with the original ground- level panorama features to obtain a robust query represen- tation. Expand
Crowdsourcing and interactive modelling for urban flood management
A novel participatory modelling and mapping approach builds on the community mapping projects across the most vulnerable wards in Dar es Salaam, Tanzania, which uses OpenStreetMap as a data platform and demonstrates its value in hydrodynamic model development and its potential for application in data scarce areas prone to urban floods. Expand
Deep Interactive Object Selection
This paper presents a novel deep-learning-based algorithm which has much better understanding of objectness and can reduce user interactions to just a few clicks and is superior to all existing interactive object selection approaches. Expand
U-Net: Convolutional Networks for Biomedical Image Segmentation
It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Expand
Using Disaster Outcomes to Validate Components of Social Vulnerability to Floods: Flood Deaths and Property Damage across the USA
Social vulnerability indicators seek to identify populations susceptible to hazards based on aggregated sociodemographic data. Vulnerability indices are rarely validated with disaster outcome data atExpand
Unbreakable: Building the Resilience of the Poor in the Face of Natural Disasters
Economic losses from natural disasters totaled 92 billion dollars in 2015. Such statements, all too commonplace, assess the severity of disasters by no other measure than the damage inflicted onExpand
DISIR: Deep Image Segmentation with Interactive Refinement
This paper presents an interactive approach for multi-class segmentation of aerial images based on a deep neural network which exploits both RGB images and annotations, and shows that user annotations are extremely rewarding: each click corrects roughly 5000 pixels. Expand