GIS unit, DALaM receives land cover mapping training using machine learning

28/09/2018

In order to build the capacity of government staff, especially GIS Unit staff at the Department of Agricultural Land and Management (DALaM). Training covering the Land Cover Mapping Hands-On was held at the GIS unit, DALaM office from September 3^rd to the7^th, 2018.

The object of this training included: 1) Strengthening the capacity for GIS Unit staff at DALaM. 2) To introduce an advanced method to conduct land cover mapping in cloud systems.

This training was run by an international consultant, Dr. Gianluca Franceschini. There were 12 people attending the workshop including 2 SAMIS national consultants and 10 from the GIS Unit staff. During the training, the trainees were introduced to SEPAL, Google Earth Engine (GEE) and ESA Hub used to access Sentinel2 imagery. Moreover, the trainees were assigned to do various exercises such as writing GEE scripts, creating segmentation by eCognition use, classification and map legend definition.

The five days of training had a many benefits and all trainees paid attention to the lessons. In addition, the meeting was interactive and all trainees was able to complete successfully all the exercises. Also, there was a multiple question and answer session between trainees and the trainer, so the time spent in the classroom was informative and had a meaningful significance. After the workshop, most of the participants got to know the basic knowledge of how cloud systems operate and consider themselves ready to proceed with the mapping exercises.

Finally, prior to the completion of the training, Mr. Saysongkham Head of the GIS unit mentioned: “In the near future if there are training opportunities in relation to GEE, we would like to attend in order to learn more about cloud systems.” The workshop was closed on 7^th of September 2018 and was deemed to be successful. Based on the programme developed during the training, DALAM assigned 3 staff to the activity. With the help of the SAMIS international and national consultants, the national cropland cover map will likely be available around September 2019