The_Problem_Statement.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&ensp;\n",
    "[Home Page](../Start_Here.ipynb)\n",
    "\n",
    "\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;\n",
    "[1]\n",
    "[2](Approach_to_the_Problem_&_Inspecting_and_Cleaning_the_Required_Data.ipynb)\n",
    "[3](Manipulation_of_Image_Data_and_Category_Determination_using_Text_Data.ipynb)\n",
    "[4](Countering_Data_Imbalance.ipynb)\n",
    "[5](Competition.ipynb)\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "[Next Notebook](Approach_to_the_Problem_&_Inspecting_and_Cleaning_the_Required_Data.ipynb)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tropical Cyclone Intensity Estimation using a Deep  Convolutional Neural Network \n",
    "\n",
    "\n",
    "*Tropical cyclone intensity estimation is a challenging task as it requires domain knowledge while extracting\n",
    "features, significant pre-processing, various sets of parameters obtained from satellites, and human intervention for analysis.*\n",
    "\n",
    "*The inconsistency of results, significant pre-processing of data, the complexity of the problem domain, and problems on generalizability are some of the issues related to intensity estimation. In this Jupyter Notebook, we run a deep convolutional neural network architecture for categorizing hurricanes based on intensity using graphics\n",
    "processing unit.*\n",
    "\n",
    "*This Jupyter Notebook is a recreation of the Research Paper titled **\" Tropical Cyclone Intensity Estimation Using a Deep Convolutional Neural Network \"** by Ritesh Pradhan, Ramazan S. Aygun, Senior Member, IEEE, Manil Maskey, Member, IEEE, Rahul Ramachandran, Senior Member, IEEE, and Daniel J. Cecil *\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Motivation for the Problem \n",
    "\n",
    "\n",
    "\n",
    "##### Since hurricanes (or tropical cyclones) possess substantial threats and cause significant damage to lives and properties, studying the stages of a hurricane is essential to determine its impact. From a scientific perspective, determining an accurate TC intensity helps. \n",
    "\n",
    "\n",
    "* More accurate historical records of TCs, mainly if a technique can be consistently applied to older satellite imagery (i.e., intensity reanalysis)  \n",
    "\n",
    "* Providing consistent intensity estimates as current intensity estimates are made via a subjective algorithm (Dvorak technique) that is applied inconsistently in different forecast areas. Initial errors are too high, especially for weak and storms that are transitioning in structure.\n",
    "\n",
    "\n",
    "In this example, we use the Saffir-Simpson Hurricane Wind Scale (SSHWS) along with intensity categorization for tropical storm and tropical depression as tropical cyclone (TC) intensity categories.\n",
    "\n",
    "\n",
    "Since TC intensity is based on maximum wind speeds (MWS), estimating the TC intensity by just using image content is a challenging problem. Several techniques utilize satellite imagery for estimating tropical cyclone intensity using Dvorak and deviation-angle variance technique (DAVT) techniques.\n",
    "\n",
    "![alt_text](images/cat+dvat.png)\n",
    "\n",
    "### Dvorak Technique : \n",
    "\n",
    "The Dvorak technique is a method using enhanced Infrared and/or visible satellite imagery to quantitatively estimate the intensity of a tropical system. Cloud patterns in satellite imagery normally show an indication of cyclogenesis before the storm reaches tropical storm intensity. Indications of continued development and/or weakening can also be found in the cloud features. Using these features, the pattern formed by the clouds of a tropical cyclone, expected systematic development, and a series of rules, an intensity analysis and forecast can be made.\n",
    "\n",
    "The primary assumption of the Dvorak method is that cyclones with similar intensity tend to have a similar pattern.\n",
    "\n",
    "\n",
    "This Figure shows some development patterns used by the Dvorak technique. Once a pattern is detected over 24 hours, the features such as length and banding from the storm are further analyzed to reach a particular\n",
    "T-number. This relates tropical cloud structures to storm intensity. Nevertheless, this technique is not perfect and still suffers from subjective biases. Due to the inherent limitations of the empirical method used, it cannot determine subtropical cyclone intensity. Today, with the successful application of the Dvorak technique for more than 30 years along with some modifications and improvements, it is used worldwide for TC intensity estimation. The Advanced Dvorak Technique provides a nearly instantaneous estimate of TC intensity objectively. It removes a large amount of the subjectivity inherent in the process and produces errors similar to a human in most cases. But despite these attempts the errors can be high during the initial phase for weak storms that are transitioning in structure. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Now that we have understood the need for building a model to eliminate human errors, let us see how to approach the problem \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## License \n",
    "\n",
    "This material is released by OpenACC-Standard.org, in collaboration with NVIDIA Corporation, under the Creative Commons Attribution 4.0 International (CC BY 4.0).\n",
    "\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "[1]\n",
    "[2](Approach_to_the_Problem_&_Inspecting_and_Cleaning_the_Required_Data.ipynb)\n",
    "[3](Manipulation_of_Image_Data_and_Category_Determination_using_Text_Data.ipynb)\n",
    "[4](Countering_Data_Imbalance.ipynb)\n",
    "[5](Competition.ipynb)\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "[Next Notebook](Approach_to_the_Problem_&_Inspecting_and_Cleaning_the_Required_Data.ipynb)\n",
    "\n",
    "\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&emsp;\n",
    "&emsp;&emsp;&emsp;&emsp;&ensp;\n",
    "[Home Page](../Start_Here.ipynb)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}