Satellites are termed an ‘all-seeing eye in the sky’. They collect data from the Earth for different applications. The principle of satellite data collection is that different objects have distinct reflectance properties, and the reflected energy is captured by satellite sensors to produce images. Three terms are commonly used to describe satellite image resolution: spatial, spectral, and temporal. Spatial resolution refers to the size of the smallest object that can be detected by the satellite sensor, e.g., 30 x 30 metres; spectral resolution is the specific wavelengths the satellite measures; and temporal resolution is the frequency of image collection or time it takes a satellite to revisit the same area, e.g., every 16 days. The data is processed/prepared for analysis by applying specialized algorithms to make it useable for research purposes. The Agricultural Research Council (ARC) generates data products derived from satellite imagery for agricultural purposes. Some of the products include maps for soil nutrients, crop yield, crop types and crop monitoring.
Soil nutrient monitoring
Soil maps are produced by collecting soil samples and analysing them in the ARC’s analytical laboratories to determine the soil nutrient contents. The macronutrients (nitrogen, phosphorus, potassium, magnesium, and calcium) and soil physiochemistry (pH and resistivity) are determined during laboratory processing. The soil data is then related to the satellite data to generate field/regional-scale maps depending on the nutrient of interest. These maps can inform soil fertility management interventions and aid in developing variable fertilizer recommendations. This application therefore contributes towards sustainable agricultural productivity.
Crop yield estimation
Current developments regarding yield estimation are focused on smallholder maize farms. Maize samples are collected, and different parameters measured on these farms to determine crop yield. The yield data is then related to satellite data with the utility of specialized machine-learning algorithms to generate yield maps. These maps are used for maize yield prediction and forecasting. This application provides an indication of the expected maize yield and aids in planning for maize shortages and surpluses to ensure sustainable maize production. Farmers can modify their crop production practices accordingly to improve their yields.
The locations of different crops are captured and used as training data to relate them to satellite imagery for crop type mapping. This application is fundamental for determining the planted area of different crops, especially on smallholder farms. These farms are challenging to map with low spatial resolution satellite data because of their small sizes and fragmented distribution which makes them difficult to detect. Thus, freely available satellite data such as Sentinel-1 and Sentinel-2 with an improved spatial resolution are used at the ARC for crop mapping on a regional scale.
Vegetation indices such as the Normalized Difference Vegetation Index (NDVI) are an indicator of crop health status. For example, areas with healthy crops will have high NDVI values while areas with sparse or unhealthy crops will have low values. This can help in mitigating factors which are adverse to optimum crop development during the season. The use of unmanned aerial vehicle (UAV) data is currently being tested for estimating crop biophysical parameters on a field scale throughout the crop growing season.
For more information:
ARC-Natural Resources and Engineering
Tel: 012 310 2692
Please share this article with your friends!
- Featured Article
- Human Resources
- Latest News
- Under cover farming