(Approx. 30 min reading incl. references)
Having discussed the importance of optical remote sensing for agriculture and its future perspectives toward 2030, we will take a closer look at optical and radar remote sensing systems, including their basic principles, advantages and limitations, with a focus on agricultural applications.
To begin with, we will review the fundamental principles of optical and radar remote sensing. Optical remote sensing relies on passive sensors that measure solar radiation reflected or transmitted by materials on the Earth’s surface within the visible (VIS), near-infrared (NIR) and shortwave infrared (SWIR) regions of the electromagnetic spectrum. In contrast, active remote sensing systems send synthetic pulses of energy and measuring the returned energy. This includes Radio Detection and Ranging (RADAR) and Light Detection and Ranging (LiDAR). Radar systems emit microwave pulses and then record the returned signal (backscatter). By measuring the travel time of the pulse, the distance of the sensor to the measured object can be estimated and the geolocation pinpointed.
The radar system used for Earth observation applications is the microwave Synthetic Aperture Radar (SAR). SAR systems overcome the distance limitations imposed by physical size of traditional radar antennas by simulating a much larger antenna, allowing these sensors to be effectively mounted on satellites. SAR instruments commonly use multispectral microwave frequency bands including the X band (3.7-2.4 cm), C band (7.5-3.7 cm), S band (15-7.5 cm) and L band (30-15 cm). Notably, radar wavelengths can penetrate media i.e. Earth surface materials due to the long wavelengths. The penetration depth varies with wavelength, deeper penetration into media occurs with longer wavelengths. In consequence, different SAR bands are suited to different applications.
A major limitation of optical RS is its dependence on solar radiation. Cloud coverage presents a problem for optical remote sensing as clouds are opaque to optical sensors and obscure the Earth’s surface from the point of view of a satellite-mounted sensor. Moreover, cloud shadows have a poor signal-to-noise ratio, which means that noise often outweighs surface signal, causing additional difficulty in data preprocessing. These effects can be mitigated through cloud masking and the creation of multi-temporal mosaics; however, cloud cover can severely reduce observation frequency in many regions of the world. Because dense time series are crucial for analyses involving vegetation phenology or near-real-time monitoring, the utility of the optical RS imagery becomes constrained under cloudy conditions.
A clear advantage of SAR systems is that they provide images regardless of daylight availability and weather conditions, transcending the limits of optical remote sensing. SAR measurements are also largely independent from cloud cover. The result is often a higher temporal frequency and a consistent acquisition schedule. SAR-based crop monitoring in near-real-time allows for precision agricultural monitoring such as yield prediction, stress detection and irrigation optimisation.
Furthermore, the unique characteristics of SAR such sensitivity to structural and dielectric (polarisation) characteristics of the surface (e. g. vegetation and soil) and media penetration allow for measuring characteristics of indicators that cannot be accurately captured by optical RS. Thus, SAR allows for improvements in agricultural applications like measuring soil moisture, assessing crops (e. g. biomass, crop height), and crop classification and identification.
However, SAR data also present notable challenges. Polarisation and multiple scattering contribute to speckle and noise. Speckle can be removed through multiple viewings; however, SAR complexity necessitates good computing capacity and can make the data more difficult to interpret. Most importantly, SAR sensors cannot capture spectral information, which is crucial for identifying important parameters and land covers such as crop type, nutrient deficiencies or changes in chlorophyll content.
SAR fusion with optical remote sensing is a promising method for addressing the limitations of each system (Khose et al., 2025). For instance, fusing remote sensing data from different sources such as optical and SAR generally performs better for cropland mapping than the use of single sources (Weiss et al., 2020). For instance, L band SAR has been used to complement optical RS and significantly improve agricultural land classification (Sicre et al., 2020).
| Optical RS | Radar RS | |
|---|---|---|
| Type of Sensor | Passive | Active |
| Spectral Range | Visible (VIS), Near-Infrared (NIR), Shortwave Infrared (SWIR), 400nm – 2500nm | Microwave, 3.7cm – 15cm |
| Information Captured | Spectral information related to surface reflectance and absorption | Structural and dielectric information related to surface roughness and moisture content |
| Atmospheric Sensitivity | Affected by clouds, aerosols and atmospheric gases (reflectance, absorption and scattering effects) | Largely unaffected by clouds, fog or aerosols |
| Data Frequency | Temporal coverage limited by cloud cover and daylight availability | Consistent temporal coverage |
| Applications in Agriculture | • Crop type identification • Vegetation health monitoring • Phenological assessments |
• Soil moisture retrieval • Crop structure assessment • Biomass estimation |
| Limitations | • Data gaps due to cloud cover • Requires atmospheric correction |
• Speckle and noise lead to high processing demands |