SDS_zoo_classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f90e7470-069d-49d4-b402-057c553b1b38",
   "metadata": {},
   "source": [
    "# CoastSeg Image Segmentation Models\n",
    "\n",
    "**Purpose**: This notebook contains an interactive dashboard that can be used to load satellite imagery and use any of the available image segmentation models to segment the satellite imagery\n",
    "\n",
    "**How it  Works**: Select the model you want to use, then select settings that will be used to run your model and then click either **Select Data Directory** or **Select Images**  to load satellite imagery for the model to run on."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9300e21d-fb60-463a-b3af-19626827ffcf",
   "metadata": {},
   "source": [
    "### Available Image Types\n",
    "\n",
    "1. **Red Green Blue (RGB)**\n",
    "   - These files are generated by default.\n",
    "\n",
    "2. **Normalized Difference Water Index (NDWI)**\n",
    "   - `NDWI = (Green - NIR) / (Green + NIR)`\n",
    "   - Uses the green and NIR bands to enhance the appearance of open water features.\n",
    "   - [Read more here](https://eos.com/make-an-analysis/ndwi/)\n",
    "\n",
    "3. **Modified Normalized Difference Water Index (MNDWI)**\n",
    "   - `MNDWI = (Green - SWIR) / (Green + SWIR)`\n",
    "   - Uses green and SWIR bands to enhance open water features and diminish built-up area features that often correlate with open water in other indices.\n",
    "   - Reduces vegetation effects better than NDWI and is more suitable for enhancing and extracting water information in regions with a background dominated by built-up land areas.\n",
    "   - [Read more here](https://www.tandfonline.com/doi/full/10.1080/01431160600589179)\n",
    "\n",
    "\n",
    "## How to Run Image Segmentation Models on the Downloaded Imagery\n",
    "\n",
    "1. **Select Imagery Type**: Choose the type of imagery you have: **RGB**, **NDWI**, or **MNDWI**.\n",
    "   - If you select **NDWI** or **MNDWI**, the **RGB** directory will automatically be converted to the selected type and saved to a new folder named after the selected type. For example, if you choose **NDWI**, a new directory called **NDWI** will appear next to the **RGB** directory.\n",
    "\n",
    "\n",
    "\n",
    "2. **Choose a Model**: Select the image segmentation model you want to use.\n",
    "\n",
    "3. **Select Performance Mode**: Choose between **Ensemble** or **Best** (Best is recommended for faster performance).\n",
    "   - **Best**: Uses only the best model available on the images. This option is faster since it involves a single model making the prediction.\n",
    "   - **Ensemble**: Utilizes all available models to create more accurate predictions. Although it takes longer to make a single prediction, the resulting segmentation is generally more accurate.\n",
    "\n",
    "\n",
    "4. **Select Imagery Directory**: Choose the directory of imagery to run the model on.\n",
    "   - Select an ROI (Region of Interest) directory under the `data` directory.\n",
    "   - Always select a directory named 'RGB'.\n",
    "\n",
    "\n",
    "## Need More Help?\n",
    "\n",
    "Check out the CoastSeg website: [CoastSeg Models](https://satelliteshorelines.github.io/CoastSeg/models/)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ac6bcc07-fddf-41ba-a2a2-082e751114b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from coastseg import coastseg_logs\n",
    "from coastseg.models_UI import UI_Models\n",
    "models_ui = UI_Models()\n",
    "models_ui.create_dashboard()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dc3ca973",
   "metadata": {},
   "outputs": [],
   "source": [
    "models_ui.model_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c204a92c",
   "metadata": {},
   "outputs": [],
   "source": [
    "models_ui.weights_directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eb41b9ae",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}