click on image for schedule in pdf format
for
Specialised Training in Machine Learning for Earth
Observation,
with Application in Agriculture
1 April to 4 July
2022
With online & in-class facilitation and field work
Ceremony and awards on 5 July 2022
Note: Closing Date for Applications is 4 March 2022
or copy and paste this link in your browser https://forms.gle/1UZ8oSrswHJiLpKT7
| From the course you can achieve: | Certificate of Completion | *Certificate of Competence | CPD points | Best Team Award | Best Student Award |
Background
Unprecedented abilities to
acquire satellite image, aerial and in-situ sensor data open new opportunities
for knowledge of the different aspects of the Earth's surface, including
activities on the ground such as agricultural practices. However, new
opportunities come with new challenges - one of these being suitable methods to
analyse and exploit such huge amounts of data. Machine Learning is considered
on option to make sense of the vast and diverse amount of Earth Observation
data. The intersection of machine learning and Earth observation remains an
emergent field, but a rapidly growing one. Machine learning, as a sub-set of
Artificial Intelligence (AI), enhances the ability of computer systems to
detect, decipher and decide from enormous amounts of data. It offers new
opportunities to break down existing barriers to socio-economic development. On
behalf of the German Federal Ministry for Economic Cooperation and Development,
GIZ implements the project of "FAIR Forward - Artificial Intelligence for all".
FAIR Forward strives for an open, inclusive and sustainable approach to AI on
an international level. The FAIR Forward's objective is to augment the
prerequisites necessary for local AI development and use across its six partner
countries (South Africa, Rwanda, Uganda, Ghana, Kenya and India). As part of
FAIR Forward's goal on capacity development, FAIR Forward has partnered with
the South African Department of Science and Innovation (DSI) and the South
African National Space Agency (SANSA) to augment capacity building on Machine
Learning for Earth Observation (ML4EO) practitioners and in doing so, support
DSI's long term strategy of capacity building, strengthening local dataset
development, building local and international networks for practitioners in the
area of ML4EO, and supporting digital entrepreneurship in the field of ML4EO.
Training provided by experts from: WITS GAES, CSIR Precision Agriculture, ARC
Training
Content
|
Module # |
Module Name |
Module Description |
Expected Outcomes |
|
Module 1 |
Introduction to Remote Sensing in Agriculture |
Introduction to concepts of precision agriculture & processes to monitor farming crop performance. Information on both present and future remote sensing technologies and satellite platforms are discussed including an introduction on the different sensor platform types (e.g. optical vs SAR vs LiDAR vs UAV applications). |
1. Understand the concepts of remote sensing and its application in precision agriculture 2. Download and process S1&2 data using SNAP 3. Display and download data via GEE |
|
Module 2 |
Classification-based remote sensing applications in Agriculture |
Desktop approach of classifying land use and land cover classes over a region of interest using spectral endmembers collected from remote sensing imagery. The classification process will also enlist a variety of ML algorithms (Spectral Angle Mapper, RF and MAXLIKE) and will involve classifier training and validation (via weighted confusion matrices). Concepts such as user’s, producer’s and overall accuracy as well as the Kappa metric will also be introduced. |
1. Classify land use/land cover and crop type using remote sensing imagery. 2. Perform ML based model training, validation and accuracy assessment 3. Test performance of different models.< |
|
Module 3 |
Field data collection for agriculture applications |
Farm excursion involving in-situ field collection and deployment of UAVs for model calibration and validation dataset creation. The exercise will also include training in the use of the different ground sensors as well as physical measurements of the crop under investigation. |
1. Exposure to field data collection that can be used for image pre-processing, ML model training, calibration and validation. 2. Be able to assemble the modelling dataset which includes the collected field data and the extracted EO dataset layers as input model predictors. |
|
Module
4 |
Machine learning modelling of important agricultural crop parameters |
This module outlines the actual predictive modelling processes for predicting continuous individual crop parameters. The concept of EO dataset upscaling for more accurate regional extrapolation will be discussed and demonstrated using multi-spectral optical UAV datasets. Concepts such as model bootstrapping, important model validation statistics and testing different ML algorithms will also be explored. R statistical software will be primarily used for the modelling and mapping process. |
1. ML based predictive modelling and mapping of important agricultural crop parameters. 2. Testing the performance ML models |
|
Module 5 |
Invited industry experts |
This stage will focus on exposure to global trends in industry and to discover how machine learning intersects the worlds of study and real-life application. |
The online seminars with select industry representatives will introduce real-life challenges which have been solved by ML |
|
Module 6 |
Mini Case Study |
Conduct a mini case study and present the findings in a joint seminar. |
Have the ability to apply ML in both qualitative and quantitative remote sensing approaches on a particular study of interest within the agriculture field. |
Training
Schedule
click on image for schedule in pdf format
|
Day in Month |
April |
May |
June |
July |
|
1 |
Fri
Module 1
|
Sun Labour Day |
Wed
Module 3
|
Fri
Module 4
|
|
2 |
Sat |
Mon Public Holiday |
Thu
Module 3
|
Sat |
|
3 |
Sun |
Tue |
Fri
Module 3 |
Sun |
|
4 |
Mon
Module 1
|
Wed
Module 1 |
Sat |
Mon Module 6 |
|
5 |
Tues
Module 1
|
Thu
Module 1 |
Sun |
Tue Closing Ceremony & Awards |
|
6 |
Wed
Module 1
|
Fri
Module 1 |
Mon
Module 4 |
|
|
7 |
Thu
Module 1
|
Sat |
Tues
Module 4 |
|
|
8 |
Fri
Module 1
|
Sun |
Wed
Module 4 |
|
|
9 |
Sat |
Mon
Module 2 |
Thu
Module 4 |
|
|
10 |
Sun |
Tues Module 2 |
Fri
Module 4 |
|
|
11 |
Mon
Module 1
|
Wed
Module 2 |
Sat |
|
|
12 |
Tues
Module 1
|
Thu
Module 2 |
Sun |
|
|
13 |
Wed
Module 1
|
Fri
Module 2 |
Mon
Module 4 |
|
|
14 |
Thu |
Sat |
Tues
Module 4 |
|
|
15 |
Fri Good Friday |
Sun |
Wed
Module 4 |
|
|
16 |
Sat |
Mon
Module 2 |
Thu Youth Day |
|
|
17 |
Sun |
Tues
Module 2 |
Fri
Module 4 |
|
|
18 |
Mon Easter Monday |
Wed
Module 2 |
Sat |
|
|
19 |
Tue |
Thu
Module 2 |
Sun |
|
|
20 |
Wed |
Fri
Module 2 |
Mon
Module 5 |
|
|
21 |
Thu |
Sat |
Tues
Module 5 |
|
|
22 |
Fri Earth Day |
Sun |
Wed
Module 5 |
|
|
23 |
Sat |
Mon
Module 2 |
Thu
Module 5 |
|
|
24 |
Sun |
Tues
Module 2 |
Fri
Module 5 |
|
|
25 |
Mon |
Wed
Module 2 |
Sat |
|
|
26 |
Tue |
Thu
Module 2 |
Sun |
|
|
27 |
Wed Freedom Day |
Fri
Module 2 |
Mon
Module 4 |
|
|
28 |
Thu |
Sat |
Tue
Module 4 |
|
|
29 |
Fri |
Sun |
Wed
Module 4 |
|
|
30 |
Sat |
Mon
Module 3 |
Thu
Module 4 |
|
|
31 |
Tues
Module 3 |
Who should apply?
Attendance, Certificates and CPD
Please be advised that participants will need to attend the full training in order to receive a certificate. The schedule is provided above and the times for training will be specified. On completion of the full training you will be awarded a certificate of completion. You could also be awarded a certificate of competence should the training receive SETA accreditation.
Professionals may be able to claim CPD points. Presently CPD points may be awarded by:
Training Venues and Logistics
Detail on venues and
logistics will be provided.
Participant approved travel and
accommodation will be sponsored by the GIZ and partners.
National guidelines on Covid-19 will be
followed.
|
Links:
This document is maintained by
Meena Lysko |
|---|