Skip to content

cadecrow/idsi-cv

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Infrastrucutre Damage Analysis from Satellite Imagery using RCNN

Abstract

In the aftermath of natural disasters, timely detection of where damages have happened and dispatching rescue teams can save thousands of lives. To determine where efforts need to be focused, a human assessment of damage is usually required. These assessments take time and money and have inherent risk. To try and solve this problem, we worked on a two-step model that would take satellite images as input, detect buildings, and determine whether or not a building is critically damaged to warrant attention.

Utilizing a ResNet50 architecture, we were able to obtain high training accuracy on damage classification, above what we expect a human can achieve. However, the accuracy on our testing set showed that improvements can be made to reduce variance. Our attempt at using L2 weights regularization is promising and given the right parameter value it is expected that the model will not overfit to the training data and perform better on the testing data. For building detection, we use Mask R-CNN (based on FPN for object detection and RESNET101).

About

Computer Vision Projct to analyze Infrastructure Damage from Satellite Imagery

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors