Use of satellite imagery is becoming widely adopted in farm operations to better guide the decision-making process of farmers. This article summarizes the basics principles of optical remote sensing applied to agricultural monitoring by describing the fundamentals, most relevant vegetation indices (VIs), and most common applications in agriculture.
Basics of Remote Acquisition
Solar electromagnetic radiation drives photosynthesis on Earth. Three things occur when the electromagnetic radiation (irradiance) hits an object. A portion of the irradiance is reflected (reflectance), a second portion is transmitted (transmittance; i.e. passes through the object), and a third portion is absorbed (absorbance) by the object (Figure 1). The proportion for each fraction (reflectance, transmittance, and absorbance) differs depending on the physical and chemical properties of the target object. Plants have a medium, low, and high reflective pattern in the green, red, and near-infrared portions of the electromagnetic spectrum, respectively. A significant portion of visible light is used as energy to trigger important biochemical processes in plants.
Figure 1. Interaction of the electromagnetic radiation and target: absorbance, transmittance, and reflectance. Infographic developed by Luciana Nieto and Ignacio Ciampitti, K-State Research and Extension.
The term “band” is used to identify the regions of electromagnetic spectrum where a satellite is sensitive to the reflected signal from the ground. Vegetation indices (VIs) relates different bands or (regions) and are used to derive biophysical information important to monitor the status of the crops. Multispectral on-board sensors retrieve data from the visible part of the spectrum (520-600 nanometers (nm)=green, 630-680 nm=red, and 450-520 nm=blue), infrared (IR), and microwaves. The IR band, depending on the characteristics of the satellite sensor, can also be divided into close IR or near IR (NIR=760-900 nm), medium IR (MIR) and far IR (FIR), or thermal (Table 1). As an example, the chlorophyll in the leaves absorbs more red (630-680 nm) and less green (520-600 nm) electromagnetic radiation. This is the reason why plants appear green to our eyes (more green radiation is reflected back to our eyes).
Figure 2. Graphical representation of the different portions of electromagnetic radiation spectrum organized by wavelength (nm=nanometer or 1 millionth of a meter). Infographic developed by Luciana Nieto and Ignacio Ciampitti, K-State Research and Extension.
Sources of Satellite Imagery
Many of the satellites orbiting the Earth have been mainly designed to monitor changes in land cover. The main characteristics of the satellites for agricultural application are:
1) Temporal Resolution indicates the frequency (time interval) for obtaining imagery data from the same point on the surface.
2) Spatial Resolution refers to the level of detail visible in an image (pixel dimensions): the smaller the area represented by each pixel in a digital image, the greater the details.
3) Spectral Resolution denotes the width of the spectral bands recorded by a sensor. The narrower these bands are, the higher the spectral resolution. Table 1 presents the number of bands per satellite sensor.
Table 1. Number of bands per satellite (different sensors).
The most commonly used sensors for agricultural applications are (Figure 3):
- MODIS (Moderate Resolution Imaging Spectroradiometer) sensor aboard the Terra and Aqua satellites, with a high temporal resolution (daily) but low spatial resolution. The minimum pixel size is 250 m, ideal for large-scale or regional work, for example county-level data. MODIS has a total of 36 bands.
- Landsat, with different mission (5, 7 and 8), has a finer spatial resolution (30 m), but with a lower temporal resolution than MODIS. The number of bands is 11.
- Sentinel 2, A and Sentinel 2 B, from the European Space Agency (twin satellites), these sensors allow more detailed spatial resolution (10 m) and weekly imagery data when both are functional. The number of bands is 13.
Figure 3. Different satellites and characterization for Spatial, Temporal, and Spectral resolution. Infographic developed by Luciana Nieto and Ignacio Ciampitti, K-State Research and Extension.
The satellite imagery collected via MODIS, Landsat and Sentinel are available to the public. Other commercial platforms are available such as Rapid Eye, a private satellite with a very high temporal (daily) and spatial resolution (5 m), but only with 5 bands available (Figure 3).
Vegetation Indices (VIs) and Applications in Agriculture
Vegetation indices (VIs) are combinations of certain spectral bands, which allow us to monitor changes in vegetation. Examples of some of the most commonly utilized indices are: Normalized Difference Vegetation Index (NDVI), Enhance Vegetation Index (EVI), Normalized Difference Water Index (NDWI), red edge NDVI (NDVIre), red edge simple ratio (SRre), and green NDVI (gNDVI).
The NDVI is universally utilized as an index for reflecting temporal and spatial differences in overall plant health and, in consequence, utilized for yield prediction. For agricultural purposes, NDVI values ranges from 0 to 1, with values ranging from 0.1 to 0.2 for soil surfaces and 0.3 to 1.0 for crop canopies. The NDVI and the greenness of an object are positively related between each other (e.g., field crop). Some VIs with their respective equations are introduced in Table 2.
Table 2. Description, acronym, equations, and references for all vegetation indices (VIs).
|
Index |
Acronym |
Equation |
Reference |
|
Enhance Vegetation Index |
EVI |
2.5*(NIR-Red)/(NIR+ 6*Red-7.5*Blue+1) |
Liu and Huete (1995) |
|
Normalized Difference Water Index |
NDWI |
(Green-NIR)/(Green+NIR) |
Gao (1996) |
|
Normalized Difference Vegetation Index |
NDVI |
(NIR-RED)/(NIR + RED) |
Rouse et al. (1994) |
|
Green Normalized Difference Vegetation Index |
NDVIG |
(RNIR-Rgreen)/ (RNIR + Rgreen) |
Gitelson et al. (1996) |
|
Red-edge Normalized Difference Vegetation Index |
NDVIre |
(RNIR-REDedge)/ (RNIR + REDedge) |
Gitelson and Merzliak (1994) |
|
Red-edge simple radio |
SRre |
RNIR/REDedge |
Gitelson and Merzliak (1994) |
Some examples of agriculture applications:
- Early-season crop classification
- Characterization of nutrient status of the plants
- Providing an overall status of field crops and other vegetation
- Prediction of biomass levels
- Forecasting crop yields; estimating crop yield before harvest
- Flooding; water excess
Summary for Satellite Data – Applications in Agriculture
Below is a selected list of the main applications of satellite data in agriculture:
- Site-Specific Management (SSM), using prescription maps to variable seeding rate and fertilization, depending on the potential of the environments within the field.
- In-seasonal (within a season) and temporal (across seasons) monitoring of crop vegetation (diagnosis of potential stress factors such as drought, nutrient deficiencies, diseases, insects, etc.).
- Forecasting crop yields at different scales: county, district, regional, state, and national levels.
- Crop scouting and sampling according to the field dimensions.
- Environmental impact assessment, fires, floods, to track land use and land cover change.
What are we expecting for the future?
- New public satellites allowing a finer time resolution (e.g. Sentinel-3) and avoiding problems with cloud interference.
- Higher spectral resolution satellites that will benefit a more intensive monitoring of functional crop growth parameters (e.g., ESA FLEX mission - planned launch date is 2022).
- More studies to focus on how to integrate information from different satellites while taking advantage of the different features from each one.
- Development of remote sensing end-to-end solutions by agricultural providers for farmers (integration with ground sensors, mobile apps, etc.).
In summary, the future of satellite data is unknown but more exciting opportunities for several agricultural applications will be available in the near future. As a final reminder, any remote sensing data (e.g., imagery collected from satellites, drones, or planes) do not replace the need for agronomists and crop scouting. These technologies provide a reliable and timely source of data to direct our efforts and increase our efficiency in targeting the main crop production problems with the ability to react and provide ‘real-time’ solutions for protecting and sustaining farming productivity.
Ignacio A. Ciampitti, Crop Production and Cropping Systems Specialist
ciampitti@ksu.edu
Luciana Nieto, KSUCROPS Dr. Ciampitti’s Lab
lnieto@ksu.edu
Rai Schwalbert, KSUCROPS Dr. Ciampitti’s Lab
rais@ksu.edu
Sebastian Varela, KSUCROPS Dr. Ciampitti’s Lab
svarela@ksu.edu
Tags: