The STARS project explores ways to use remote sensing technology to improve agricultural practices of smallholder farmers in sub-Saharan Africa and South Asia with the aim to advance their livelihoods. "STARS" stands for Spurring a Transformation for Agriculture through Remote Sensing.
“STARS” stands for Spurring a Transformation for Agriculture through Remote Sensing. It is an international project led by the University of Twente in the Netherlands and funded by the Bill and Melinda Gates Foundation. The project investigates how very high spatial resolution satellite images of half a metre to five metres pixels and those derived from unmanned aerial vehicles (UAVs) can be used to monitor and map smallholder farms.
Camera-carrying UAVs were chosen for this project because of their unprecedented centimetre-level spatial resolution and because they can be flown at low altitudes below the cloud level. Using UAV technology, a temporally dense time series of images can be constructed that allow to closely track crop changes over time. Also, on-board cameras can collect images spectrally compatible with those provided by earth observation satellites, making it possible to conduct multi-scale analyses.
The STARS project focuses on three regions: West Africa (Mali and Nigeria), East Africa (Tanzania and Uganda) and Bangladesh in South Asia. These regions present three highly relevant agricultural problems: land tenure and field performance (West Africa), agricultural statistics and food security at regional and national levels (East Africa), and irrigation scheduling (Bangladesh).
Land tenure registry in West Africa
In West Africa, the team is led by ICRISAT (International Crops Research Institute for the Semi-Arid Tropics). They own two fixed-wing UAVs. These UAVs can be used to obtain both true colour (red, green and blue or RGB) and near-infrared images. These images support the identification and mapping of agricultural plots. More accurate and transparent land tenure information could help West African smallholder families to secure their land use rights.
In parallel to the collection of UAV images, a field team worked on the ground to measure and map several farm management units. Later, this geographical information was combined with the information gathered by interviewing farmers to create a land tenure registry.
Moreover, by using photogrammetric techniques, the STARS team was able to create precise digital elevation models of the photographed areas. These elevation models can be used to support land management and to derive crop height — a valuable metric for studying crop status that can be used together with textural and spectral information.
The West African team also owns an eight-armed helicopter model (octocopter) that can carry an RGB and a multispectral (5-bands) camera. This UAV is used in a more experimental fashion, as researchers attempt to determine whether different crop varieties can be identified from multispectral images, or if certain measures of crop health, such as leaf area index or chlorophyll content, can be accurately derived from UAV images.
Supporting food security policy-making in East Africa
The East African team, led by the University of Maryland in the US, supports the collection of national agricultural statistics and food security policy-making in Tanzania. In this case, the team used two fixed-wing UAVs to map maize-based agricultural systems.
After completing the UAV flights, results were scaled up to the national level, using satellite data and crowd-sourced information from the ground. This resulted in a cropland map that was shared with officials at the Tanzanian Ministry of Agriculture. Maps like these, if created in a timely fashion, can help agricultural experts more accurately forecast yields at a national level, and to make informed decisions about the state of food security.
The East African team, like the West African team, also operates an octocopter. This UAV is used to perform in-farm experiments aimed at better understanding how multispectral UAV images can be used for mapping cropping systems and their condition.
Optimising irrigation scheduling in Bangladesh
The Bangladeshi team, led by CIMMYT (International Maize and Wheat Improvement Centre), uses two octocopters. One of them is equipped with the same model of RGB and multispectral camera that is used by the African teams. However, the spectral bands chosen by the Bangladeshi team are slightly different so as to allow for a more precise characterization of crop photosynthetic activity. This UAV is also used to map how the fraction of vegetation cover changes over time. This information is key for optimising irrigation scheduling.
The second octocopter is equipped with a thermal camera that was used to assess canopy temperature, which is key for designing an improved irrigation strategy. Although Bangladesh is rich in water, cereal farmers must pump surface water to cultivate their lands during the dry months of winter. The STARS project hopes to help Bangladeshi farmers to grow an extra crop per year to improve their financial and food security. Remote sensing technology helps farmers to optimise the use of water pumps and it provides valuable information for a sustainable intensification of their lands.
To undertake the agricultural analysis, the STARS team has had to overcome several UAV-related challenges. The team had to secure permission to fly from relevant authorities and had to train various local staff members so that flights could be performed in a safe and timely fashion. Team members also had to inform local people of the activities and engage with farmers in the data collection process.
Operating the UAVs twice per month over each field was challenging. The teams had to arrange complex field visit logistics. They only could fly the UAVs during the relatively brief period of time of the day when environmental conditions are optimal for gathering aerial imagery. Although all regional teams did an excellent job, there were a few crashes. Some of the UAV cameras and batteries overheated and did not function properly.
Another challenge was related to figuring out how to transfer the UAV data from the field to the central offices of the regional teams. This was a necessary task, as high computational power and specialised software is required to process the several gigabytes of information that are collected during a typical UAV flight campaign.
Field data is needed to calibrate the retrieval of crop properties and to classify aerial images collected by UAVs. Hence, the STARS team UAV usage was accompanied by intensive field campaigns, where a wide array of crop-specific information and measurements were collected, such as leaf area index using smart phones. The teams also collected ground control points, which are accurately surveyed geographic coordinates needed to properly geo-reference UAV imagery. Geo-referenced UAV imagery matches up with other spatial data, and can be combined with other spatial datasets in Geographical Information System (GIS) and remote sensing software.
Finally, there is the challenge of calibrating UAV images. Calibration is needed to ensure that the quality of the images is as high as possible so that multi-temporal and multi-scale analyses can be done. However, image calibration remains challenging and STARS researchers are still investigating ways to operationalise it.
Despite these challenges, the STARS project is steadily progressing towards its goal of determining how remote sensing technology can be used to shed light on the complex agricultural systems in which smallholder famers operate. As such, STARS is the first stepping stone on a path that leads to more sustainable agricultural production in emerging economies.
The STARS Project web site