Exploring the use of earth observation and data science for agricultural statistics to complement the census dataset: Case study for Namibia Statistics Agency

(Published in Statistical Journal of the IAOS 36 (2020) S121–S136)

By Tulimegameno Amutenya

Published on 24 February 2021

Abstract.

Agriculture is the backbone of human life, it enables for food security, health and economy. Yet, many countries in Africa suffer from poor accessibility to agriculture data which is crucial for policy makers and farmers. Half of Namibia’s population depend on agricultural activities, for as their main income source, much of which is undertaken on smallholdings.

Therefore, compiling statistics around agricultural outputs is of primary concern to many national statistics agencies Unfortunately, challenges to account for agriculture crop production statistics include low frequency of data collection, lengthy data processing periods, and the lack of timely output which can be linked to policies and decision making.

This paper explores the use of satellite imagery and data science techniques in a statistics agency to complement the agriculture census. The paper assessed Google Earth Engine for image processing and extracted a range of indices (NDVI, SAVI, MSAVI and GLCM and Tasseled Cap Index based) in order to identify smallholder farmers’ plots and estimate the field area in a rural village in Namibia.

Although groundtruth data was not available at the time of this issue, the findings showed a promising starting point for a scaled project.

The full paper is available to download here.