rahul263-stack/fire_prediction

This project is aimed to build a model that can predict fire and smoke...

Forest Fire Prediction

Motivation

Forest Fires cause severe hazard in endangering vegetation, Animal and human life across the world, apart from this it can be a factor for Airborne hazards , Water pollution and other Post- fire risks. Also after Post-fire is very hard to control and even wildland fire fighters face several life-threatening hazards including heat stress, fatigue, smoke and dust, as well as the risk of other injuries such as burns, cuts and scrapes, animal bites, and even rhabdomyolysis. Between 2000–2016, more than 350 wildland firefighters died on-duty. Only Amazon forest fire cost Brazil US$957 billion to US$3.5 trillion over a 30-year period.

CNN Based Early stage detection classifier

The working model is deployed here

File Structure

FIRE_PREDICTION
├───data
│   ├───fire
│   └───Non_fire
├───model
│   ├───fire_detection.model
│   │   └───variables
│   └───sample_classifier_model  ## classic model 
├───readme_files  ## extra files for the readme.md 
├───techniques    ## source code 

How to setup

follow these simple steps to setup the environment

Step 1: make sure you have python installed version 3.5 to 3.7 (anyone) 
Step 2 : make a virtual environment( from my prefrence i use miniconda )
Step 3 : use this command to automatically download each required package :
                    pip install -r requirements.txt 

if everything didnot go as planned use this command 
                    pip install -r requirements_old.txt

step 4 : open the directory and run the app by :
                    streamlit run app.py 

step 5 : everything should be in order, and tryout my app :
    

Tools used

Thanks

@sabstian thrun
@ udacity
@ Lovely Professional University

license

the license is MIT but it would be better if you include me(karikeyshaurya) also