-
Language: Python​
-
Libraries: OpenCV, Keras, Scikit-learn, Scikit-fuzzy, Scikit-image, NumPy, TensorFlow, Matplotlib, Plotly
-
Platform: Python file
Flood Detection
Purpose
Detect flood and estimate the extent of flood from crowd-sourced mobile images of the disaster affected area, using fuzzy based semantic scene analysis.
Models
CRFasRNN
Dataset
Large number of crowd sourced and random flood scene images are collected to form our dataset. A total of 320 images are tested, out of which 180 are flood scene images and 140 are non flood images.
Tasks Done
Image Processing
Color based Segmentation
Largest Contour Segmentation
Face Detection
Vehicle Segmentation
Fuzzy rules
Module Integration
Applications
Disaster Management - Flood estimation can help the government prioritize the areas to where immediate help should be provided.