Geospatial Technologies Driving Digitalization of Agriculture: Innovations, Trends, and Future
Agriculture is one of the most important sectors of the Indian economy. It contributes to approximately 17% of the total GDP and roughly 60% of the total population is dependent on this sector for its livelihood. To ensure continuous and sustained growth in this sector, integrated “Digital Agriculture” is being used by many companies, government bodies and organizations.
Digital Agriculture comprises using digital technologies for crop monitoring, livestock management and other processes in agriculture. Its basic aim is to increase production of the crops with sustainable development and reduce the risk of crop loss or damage.
Geospatial technologies have an integral part to play in the digitalization of agriculture, being used by different organizations across India to monitor individual agricultural fields due to their efficacy in monitoring different aspects of crop cycles as well as forecasting yield and production estimates.
Using Geospatial Technologies to Improve Agricultural Processes
Different stages of crops can be seen using Remote Sensing imagery (Satellite as well as aerial and UAV) which helps users analyze the crops. Various agencies like ESA, USGS and projects like the EnMap Mission provide open-source imagery captured by SAR, multispectral, and hyperspectral satellites, which have tremendous applications in Crop health monitoring and assessment.
As far as the process of Remote Sensing based crop monitoring is concerned, the very first step is to choose the correct imagery. Correct means the imagery captured on (or near to) the date of sowing of crops. This selection of imagery is very important as the pixel values of the images depict the stage-specific reflectance or backscatter values. Preprocessing of the imagery is also important to minimize the noise.
To get the actual ground data, extensive ground surveys are required to train the models for crop-type classification. After the classification of different crops sown in the study area, the cadastral boundaries are overlaid to get the micro-level picture. Due to the extensive use of Machine and Deep Learning algorithms, the accuracy and efficacy of the outputs have increased.
As optical images are marred by clouds or haze in the rainy season, SAR plays a complementary role due to its penetration capability of the cloud. The basic hypothesis in the Remote Sensing based crop phenology assessment is that the temporal profile of backscatter values mimics the crop growth and senescence over a cropping season. So, this crop cycle can be simulated in terms of backscatter values using mathematical models (For instance, the Badhwar Rice growth model and Tor Vergata model) to get the parameters like Crop emergence date, harvesting date, peak biomass and so on.
As crop health depends on multiple factors like management practices, rainfall, Soil Moisture, fertilizers etc., continuous monitoring of the crops has become very necessary. Mobile apps, IoT based sensors are being used extensively to measure different parameters like Soil Moisture, humidity, temperature, pH, nitrogen content and so on. Temporal modelling of these parameters along with reflectance/backscattered values gives more detailed information related to crop growth.
Geospatial technologies are also used to model crop yields. Biophysical parameters, especially biomass and their integration with biochemical parameters are of paramount importance. Various crop simulation models having agronomy parameters have been used by many researchers to forecast yield estimates. Highly complex deep learning models are currently being used with SAR data for accurate yield estimation.
Current Trends in Digital Agriculture
The creation of mobile apps, web APIs and the use of IoT-based sensors and drone imagery are the current trends. A huge amount of data is being captured, produced, and processed by different companies and organizations by using the cloud. Mobile apps are useful to generate agriculture field-level advisories for individual farmers.
Farmers upload photos of crops in the app and the robust models used in that app process these photos and shoot the results pertaining to the field. This analysis helps farmers take prompt action in case of any crop disease. Drone imagery is extensively being used to map damaged agricultural areas due to floods or any other calamity. This helps the policymakers and ministers in the agriculture department to devise efficient policies (like Pradhan Mantri Fasal Bima Yojana). Many private and nationalized banks use Remote Sensing based estimates to settle crop insurance-based claims.
A recent survey shows that banks like HDFC and insurance companies like Bajaj Allianz are leading the way in Remote Sensing-based insurance claim settlements. This trend will continue due to the availability of high-resolution imagery in the recent future, making crop health monitoring more accurate. Integration of SAR and hyperspectral (data fusion) imagery is another aspect which will be exploited by the Geospatial data scientists.
Further, it will facilitate the integration of biophysical and biochemical parameters which is the crux of remote sensing-based agricultural studies. As climate change is inevitable, climate-resilient agricultural practices will get momentum. Weather modelling, Numerical weather prediction, climate risk assessment modelling and their use in the dissemination of advisories constitutes the major work of a few companies/organizations in India, like the Indian Meteorological Department. Such developments are also necessary to investigate fraudulent practices (if any) to ensure accountability and transparency in the bureaucracy.
New Technologies and Their Role
Data Science, Machine Learning, AI, and Cloud computing are the buzzwords nowadays. These are very useful technologies, and they are indispensable when it comes to the generation of efficient Decision Support Systems. At the same time, the domain experts in Geospatial technologies are equally important. If accurate information from ground surveys is available, training and testing of the Machine and Deep Learning models become easier.
Hence, many companies are focusing on such surveys. Actual biomass measurement by using destructive testing methods is being done by many organizations like, ISRO, MNCFC and others and complex simulation models like Radiative Transfer Model, Water Cloud Model and so on are being used to forecast yield.
These models can be replaced by Deep Learning models but for such replacement, domain knowledge in Remote Sensing and agronomy is very important. To make each farmer techno-savvy and smart, capacity-building programs are being designed to ensure that the farmer explores all the possible resources available for his agriculture.
Conclusion
Digital agriculture has evolved much since its inception, and it is still growing. In India, remote sensing-based agriculture monitoring is very challenging given the diverse agro-climatic zones and agricultural practices. To ensure that small landholding farmers are getting useful insights pertaining to their cropped fields, the exploitation of Geospatial technologies is tantamount to policy making and implementation.
As far as policymaking is concerned, the government has mandated that companies and organizations working on agriculture projects use Geospatial technologies for their analyses. This has given rise to many start-ups and Technology Incubation Centers which are using huge amounts of Remote Sensing and GIS data for their projects.
In a nutshell, it can be stated that Geospatial technologies are the main drivers of Digital agriculture and in the recent future, their usage will increase exponentially.
Author: Satej Panditrao, Technical Manager, Association of Geospatial Industries (AGI)
Leave a Comment