EtienneDavid / Learn_Data_Science_For_Agriculture

How to learn Data Science for Agriculture ?

Practical Deep Learning fro agriculture slides

You can find my slides for Deep Learning training here: https://1drv.ms/f/s!AjR7TH0MGCgIhY162GFktgQ45Rl11g

Learn_Data_Science_For_Agriculture

You want to use the power of Data Science for Agriculture ? I've gathered ressources to improve in both domains in order to be able to tackle the biggest issues

I- The biggest challenges to adress

A healther, more, sustaniable world achieved through imagination, innovation and collaboration

Grand challenges:

• Increase production of more nutritional food
• Presere naturalressources like water and topsoil
• Develop sustainable sources of bioenergy

Innovative Technology to accelerate discovery

• Genomics / Phenomics
• Crisp genome editing
• Robotics / Imaging
• Computation

Impact

• Licensed technology
• Nex company formation
• Improved crops
• Next-generation scientist

II- Back to farm: What you need to understand on farming

Plant Science

• Photosynthesis
• Genomics and phenomics
• Drought tolerance
• Microbial interactions
• Response to pathogens

Agriculture

• Agricultural history
• Nutrient uptake
• Soil Management
• Water management

III- Deep Learning

First steps in Python

• Learn to use Python for common tasks: https://automatetheboringstuff.com/
• Computer vision
  • https://www.pyimagesearch.com/
  • https://eu.udacity.com/course/introduction-to-computer-vision--ud810
• Good practices: https://docs.python-guide.org/

Machine Learning and Deep Learning

• Stanford syllabus: http://web.stanford.edu/class/cs20si/syllabus.html
• fastai.com : a really good introduction for coders to Deep Learning
• Siraj Raval Maths of Intelligence: https://www.youtube.com/watch?v=xRJCOz3AfYY&list=PL2-dafEMk2A7mu0bSksCGMJEmeddU_H4D
• Deep Learning, by Ian Goodfellow, more theorical: https://github.com/janishar/mit-deep-learning-book-pdf

Phenotyping

Classification

• Phenology detection

Object Detection

• Common workflow:
  • https://medium.com/@karol_majek/10-simple-steps-to-tensorflow-object-detection-api-aa2e9b96dc94
• Plants counting
  • Big plants
  • Wheat
• Organ counting
  • Ears: https://www.sciencedirect.com/science/article/pii/S016819231830337X

Segmentation

• Green Fraction and LAI

Yield prediction

• DSatellite: Deep Learning for yield prediction: http://sustain.stanford.edu/crop-yield-analysis/

Resources

Papers

• https://github.com/simonMadec/awesome-phenotyping-deep-learning
• https://github.com/gjy3035/Awesome-Crowd-Counting

Projects to reproduce

• https://www.kaggle.com/c/plant-seedlings-classification

Data scientist environment

• Anaconda
  • Tensorflow & Keras
  • Pytorch
• Jupyter: Prototyping
• Visual Studio
  • Debugging feature
  • Testing functions

Other

• Data Science for Agriculture network in France: http://www.modelia.org/moodle/

Inspirations

I've been largely inspired by the work of Siraj raval ( ||Source|| on github) and Thomas Roca for making this curriculum, and try to share knowledge on Data Science and Agriculture.