IntelligentNetworkSolutions / WasteDetection

Prototype for WasteDetection using high quality Satellite Imagery

WasteDetection

Prototype for WasteDetection using high quality Satellite Imagery

Set Up

Python 3.9

• Install Python

  • Download Python (python.org)
  • Launch python-3.9.13-amd64.exe
    • Install Python 3.9.13
      • Choose Customize Installation
    • Optional Features
      • Check
        • pip
        • py launcher
        • for all users (requires elevation)
      • Next
    • Advanced Options
      • Check
        • Install for all users
        • Associate files with Python (requires the py launcher)
        • Add Python to enviroment variables
        • Precompile standard library
      • Customize install location
        • C:\Program Files\Python39
      • Install

• Install numpy

  • Launch cmd
    • Enter the following command:
      • pip install numpy

Orfeo Toolbox

• Download – Orfeo ToolBox (orfeo-toolbox.org)

• Extract the downloaded file directly to the C: drive

• You must have “Visual C++ Redistributable for Visual Studio 2019” installed for using this package.

  • It can be downloaded freely from microsoft

GDAL

• Download at (gisinternals.com)

  • Download gdal-3.6.3-1930-x64-core.msi
  • Download GDAL-3.6.3.win-amd64-py3.9.msi

• Installation

  • Launch gdal-3.6.3-1930-x64-core.msi
  • Choose the Complete Setup Type
  • Install for all users
  • Select Python Installations
    • Python 3.9 from registry
  • Install location should be C:\Program Files\GDAL
  • Launch GDAL-3.6.3.win-amd64-py3.9.msi

• System Enviroment Variables

  • Edit variable PATH
    • New C:\Program Files\GDAL\
  • Add variable GDAL_DATA
    • C:\Program Files\GDAL\gdal-data
  • Add variable GDAL_DRIVER_PATH
    • C:\Program Files\GDAL\gdalplugins
  • Add variable PROJ_LIB
    • C:\Program Files\GDAL\projlib

Visual Studio 2022

• Download (visualstudio.microsoft.com)

• Install

  • Check
    • ASP.NET and web development
  • Installation details (on the right)
    • .Net Framework 4.8 development tools
    • Entity Framework 6 tools
    • IntelliCode

• Launch Visual Studio 2022

• Clone the Repository Waste Detection

• Clone directly to C:/

• • C:\WasteDetection

• Restore Nuget Packages

• Build Solution

• Update appsettings.json

  • OrfeoToolboxPath
  • • C:\OTB-8.1.1-Win64
  • GDALToolsExesPath
  • • GDALToolsBatsPath
  • CmdPath
  • • C:\Windows\System32\cmd.exe

• Launch using IISExpress

• Thrust SSl Certificate from VS22

• Download input data to use in prototype Download Input Data

• Extract the data in the wwwroot\detection\prepared_inputs directory of the Solution

  • create the detection if it is not already there
  • Example full path of the input image used in the Detection process:
    • C:\Visual Studio Projects\WasteDetection\WasteDetection\wwwroot\detection\prepared_inputs\1to10.tif

Contents

Data used

• All inputs are full paths

• Input Image (full process, calculate image statistics, train image classifier, image classification):

  • ...\detection\prepared_inputs\1to10.tif

• Input Statistics:

  • ...\prepared_inputs\statistics\1to10.xml

• Input Model:

  • ...\prepared_inputs\models\model_1to10.mdl

• Input Training Layer:

  • ...\detection\prepared_inputs\training_layers\training_classes.shp

• Input Control Layer:

  • ...\detection\prepared_inputs\control_layers\control_classes.shp

• Input for Raster Calculation:

  • Input is Output from Image Classification

• Input for Sieve:

  • Input is Output from Raster Calculation

• Input for Polygonize:

  • Input is Output from Sieve

Map

• An Open Layers Map
  • zoom in
  • zoom out
  • drag to move
  • switch layers

• Layers
  • Base Maps
    • osm
      • Open Street Map
      • default visibility: false
    • input_image
      • The starting input image.
      • • The image on which the model was trained on and on which the image classification was done
      • default visibility: true
  • Vector Layers
    • input_training_vector
      • The input training classes as polygons (Vector layer .geojson)
      • color: orange
      • default visibility: false
    • input_validation_vector
      • The input validation classes as polygons (Vector layer .geojson)
      • color: yellow
      • default visibility: false
    • output_vectorized_result
      • The detected waste as polygons (Vector layer .geojson)
      • color: red
      • default visibility: true

The data shown comes from the dowloaded input

Detection

• Run the full waste detection process on the downloaded input image
• • C:\WasteDetection\WasteDetection\wwwroot\detection\prepared_inputs\1to10.gif
• The other needed parameter are drawn from prepared inputs folder (downloaded input data)

Steps

• Allows the option to run each step of the waste detection process individually
• The output of each step can be used in the following step, only the path is needed
• The Images generated by the steps: ImageClassification, Raster Calculator and Sieve can be viewed correctly in the QGIS software
• Since this is a prototype some output paths might not be shown, but there is a high likelyhood that their output will be generated in the appropriate folder inside wwwroot\detection\step_name\calculated in this case these path can be constructed manually and still be used as input to the next step

Licences

• Dependency Names
• Dependency GitHub Repositories
• Dependency Licence Files