PlanetScope in Practice: Monitoring Crops Across Smallholder Drylands in India
Region Asia
Summary
In dryland regions where rabi crops are vital for survival, lack of data can lead to a gap between agricultural policies and the actual needs of farmers. Integrating high-frequency satellite imagery allows organizations like the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) to track growth patterns across millions of hectares without the delay of manual reporting. This information provides a reliable foundation for national food security to help officials understand exactly where food is being grown. As a result, local governments can make better decisions regarding procurement programs and disaster response to protect rural livelihoods.
Decision-Making Context and Needs
Drylands are home to more than two billion people worldwide, roughly half of whom are smallholder farmers living in poverty. Many depend on crops like millet, sorghum, chickpea, and groundnut for their livelihoods.
These crops are climate resilient and important for local nutrition. However, dryland farming systems are complex. Fields are often small and fragmented, weather conditions can shift quickly, and reliable crop data is limited.
Accurate monitoring of crop production is a challenge for governments, making effective planning difficult. Gaps in reliable data can influence agricultural policies, affecting subsidy distribution, processing of procurement programs, and the speed of government response to crop losses caused by disasters.
Crop monitoring has traditionally relied on field surveys. These surveys provide valuable information but they take time, require significant resources, and are often limited in coverage.
In dryland regions with smallholder farms, plots can be scattered and hard to reach and surveys may overlook details about which crops are being grown and where. At the same time, accurate data on cultivated areas is crucial for planning national food production and ensuring reliable supply.
In India, ICRISAT conducts agricultural research for development across the drylands of Asia and sub-Saharan Africa. For more than 50 years, the organization has worked to reduce poverty, hunger, and malnutrition through research and innovation.
In one of ICRISAT’s research studies, they leveraged Planet monitoring data and combined it with spectral matching technique (SMT) to improve crop area assessment.
Value Chain Diagram
Decide
- Improve agricultural planning by enhancing national production estimates and supply chain reliability.
- Support policy implementation through evidence-based decisions for subsidies and procurement programs.
- Enable rapid response with the data foundation required for timely disaster relief and government intervention.
Apply
- Spectral Matching Techniques (SMT): use temporal NDVI signatures and growth stages to distinguish specific crop types.
- Cropping System Tracking: monitor land use, crop stress, and rabi season patterns across scattered landscapes.
- In-season Monitoring: provide continual insights to bridge the gap between field surveys and administrative reports.
Data production
- Crop classification generates high-accuracy data identifying crops such as wheat (91% accuracy) and mustard (94% accuracy).
- Cultivated area assessments provide detailed land-use data, maintaining a 96% correlation with official district-level statistics.
- Temporal NDVI signatures deliver time-series data at seven-day intervals to track vegetation health and growth cycles.
Sense
- PlanetScope leverages 3.7-meter medium-resolution imagery for high spatial detail in smallholder plots.
- Near-daily revisit of area of interest (AOI) ensures persistent monitoring.
- Multi-spectral observations capture the spectral detail necessary for calculating vegetation indices and crop-specific signatures.
Earth Observation Data Products and Information
How Daily Satellite Imagery Supports Crop Monitoring | Planet + ICRISAT
Satellite monitoring provides an efficient way to monitor crops at scale, offering continual coverage, frequent observations, and accurate measurements across large areas.
Although there are several satellite data providers that collect agricultural data, their spatial resolution is sometimes not detailed enough to detect changes in small and fragmented fields.
Very high resolution commercial satellites can capture more detail, but they are often expensive and usually cover smaller areas.
According to ICRISAT’s Principal Scientist & Cluster Leader, Geospatial Sciences and Big Data, Dr. Murali Krishna Gumma, “We’ve found that PlanetScope strikes the right balance. With 3.7-metre medium resolution and near-daily global coverage, it provides both the spectral detail and the revisit frequency we need. This makes it unrivalled for monitoring crops across broad dryland regions, and crucially, for detecting cropping patterns even in smallholder landscapes.”
Underlying Observations and Data Sources
ICRISAT’s research focused on three districts in India that are characterized by small plot farming. The goal was to assess how satellite data could accurately map cultivated areas and cropping patterns during the Rabi, or winter, season.
The Rabi season was selected because skies are usually clear. This allows satellites to capture consistent imagery, unlike the monsoon season when cloud cover often limits observations.
ICRISAT researchers used 3 meter PlanetScope imagery collected at seven day intervals. Using PlanetScope and SMT, researchers were able to map land use and crop cover, monitor crop stress, track cropping systems, and support in season crop monitoring.
SMT uses the relationship between crop growth stages and temporal NDVI signatures. This makes it possible to distinguish different crop types and track cropping patterns over time.
Impacts and Benefits
The study achieved an overall classification accuracy of 94%. Accuracy for specific crops reached 91% for wheat and 94% for mustard.
Chickpea fields were also identified successfully, although fragmented plots slightly reduced classification accuracy.
Some misclassification occurred in areas where irrigated wheat and mustard had similar characteristics. Mustard plots partly covered by trees also created challenges.
“When we compared our results with official district-level statistics, the correlation was 96%—showing strong agreement with trusted government sources,” added Dr. Murali.
These findings demonstrate that PlanetScope combined with SMT can provide reliable crop area assessments across large dryland regions.