ials / percepcion2

Percepción Remota Avanzada

UNAL - 2022 - Segundo Período Académico

Este es un curso que se dictará entre el 10 de agosto de 2022 y el 30 de noviembre de 2022 (miércoles entre 2:00 pm y 6:00 pm)

PARTE 1: RADAR DE APERTURA SINTETICA (SAR)

Semana 1 - 8 a 12 de agosto

• Lecture:

• Repaso de Radar (Heather McNairn, 2021)

• Introduction to SAR Data (Kristenson, 2020)

• Readings:

  • Introduction to SAR
  • A tutorial on Synthetic Aperture Radar (Moreira et al., 2013)

Semana 2 – 15 a 19 de agosto

• Lecture:

  • Impact of Soil and Crop Characteristics on SAR Response
  • Introduction to ESA Toolboxes (Fitrzyk, 2019)

• Practical:

  • Crop monitoring with Sentinel-1 data (SNAP Practical):
    • General Guide
    • Video by Amalia Castro
    • Step-by-Step Procedure
    • Exercise Dataset
      • Wicket
      • ESA 2019 course

• Students define data & application for their first report: "SAR image analysis"

• Readings:

  • Urban objects detection from C‑band SAR satellite images

Semana 3 - 22 a 26 de agosto

• Lecture:

  • SAR Physics (CNES, 2018)
  • Synthetic Aperture Radar (SAR) & Google Earth Engine

• Home activities:

  • Working with SAR data in Google Earth Engine
  • Students start writing their first report

• Readings:

  • Principles, Data Access, and Basic Processing Techniques (Meyer, 2019).

Semana 4 - 29 de agosto a 2 de septiembre

• A Python package for interactive mapping with GEE

  • Students start writing their first report: "SAR image analysis"

• Additional resources:

  • OpenSAR Toolkit

• Additional references:

  • A Layman’s Interpretation Guide to L-band and C-band Synthetic Aperture Radar data (CEOS, 2018)

Semana 5 - 5 a 9 de septiembre

• Analysis Ready Sentinel 1 Data in GEE

• Home activities:

  • SAR image analysis
    • Students conduct SAR image processing tasks for their first report

Semana 6 – 12 a 16 de septiembre

• What is Open Data Cube?

  • CEOS ODC - Lessons Learned

• How to use the ODC-Google Sandbox

• Home activities:

  • SAR image analysis
    • Students conduct SAR image analysis tasks for their first report

Semana 7 - 19 a 23 de septiembre

Examen No. 1 – 21 de septiembre

• Notas Examen No. 1

• Home activities:

  • SAR image analysis
    • Students revise their first report

Semana 8 - 26 a 30 de septiembre

• Lecture:
  • Machine Learning for Remote Sensing
  • A Walk-Through on Machine Learning Techniques for Sentinel Big Data Fusion

Entrega Informe No. 1 – 2 de octubre

PARTE 2: DEEP LEARNING FOR IMAGE ANALYSIS

Semana 9 - 3 a 7 de octubre

• Lecture:

  • Foundations of Deep Learning
  • Building a linear regression model

• Practical:

  • Hands on DL foundations (see the slides and do the work outlined there)

• Home activities:

  • Read Mechanics of learning
  • Read and create Colab notebook on Fully connected networks
    • Slides

Semana 10 – 10 a 14 de octubre

• Lecture:

  • Deep Learning Basics
  • Convolutional neural networks

• Practical:

  • Convolutional neural networks (see this document and write the code included there)

• Home activities:

  • Complete your Colab notebook on Convolutional neural networks

Semana 11 – 17 a 21 de octubre

• Lecture:

  • Image segmentation

• Practical:

  • Download nuclei data from here and upload it to your GDrive
  • Check idlmam.py and upload it to your GDrive
  • Replicate code for image segmentation

• Home activities:

  • Complete your Colab notebook on Image segmentation
  • Read Beyond RGB: Urban Remote Sensing With Multimodal Deep Networks (Audebert,2018)

Semana 12 – 24 a 28 de octubre

• Topics:

  • Deep Learning for multispectral images
    • Deep Learning for Remote Sensing

• Practical:

  • Semantic segmentation of aerial images with deep networks
    • DeepNetsForEO.ipynb

• Homework:

  • Students adapt the Le Saux's notebook for their second report
  • Readings:
    • Le Saux's presentation
    • Le Saux's website

Semana 13 – 31 de octubre a 4 de noviembre

Topics:

• Semantic segmentation revisited

• Deep Learning for Earth Sciences (Camps-Walls et al., 2021)

  • Introductory chapter

• Readings:

  • Object Detection and Image Segmentation with Deep Learning on Earth Observation Data: A Review-Part I: Evolution and Recent Trends

• Homework:

  • Students start writing their second report: "Image analysis using DL"

Semana 14 – 7 a 11 de noviembre

• Vegetation Remote Sensing

  • Deep Learning for Automatic Outlining Agricultural Parcels
  • Vegetation detection using DL
  • Plant diseases recognition using DL - a review

• Practical:

  • Students adapt the Le Saux's notebook for their second report

• Readings:

  • U-Net feature fusion for multi-class semantic segmentation

Homework:

• Students keep writing their second report: "Image analysis using DL"

Semana 15 – 14 a 18 de noviembre

• Topics:

  • Deep Learning for Water Resurces

    • RivaMap: An Automated River Analysis and Mapping Engine
    • DeepWaterMap

  • Vision Transformers for Semantic Segmentation

• Practical:

  • Inference of water from multispectral images

• Additional resources:

  • Canopy/land cover analysis for the City of Sammamish

• Readings:

  • Object Detection and Image Segmentation with Deep Learning on Earth Observation Data: A Review-Part II: Applications
  • A survey of public datasets for computer vision tasks in precision agriculture

• Homework:

  • Students write their second report: "Image analysis using DL"

Semana 16 – 21 a 25 de noviembre

Examen No. 2 – 23 de noviembre

• Notas Examen No.2

• Example of ML for image analysis:

  • Crop disease severity

• Homework:

  • Students complete and revise their second report: "Image analysis using DL"

Semana 17 - 28 de noviembre a 2 de diciembre

Entrega Informe No. 2 – 30 de noviembre

Notas Informe No. 2