AI Based Yield Prediction Engine
Grundo's in house yield prediction engine for various staple crops is an essential data source for financial market decisions, policy assessments and ground information verification. The first system version is currently deployed for our European customers and undergoing rigorous accuracy benchmarking.
Figure Prototype Yield Prediction tailored for Switzerland.
Artificial Intelligence Crop Classification
We combine agriculture and sustainability subject matter expertise with computational methods to provide fully customizable crop- and plant-detection capabilities. Our artificial intelligence algorithm for crop classification uses neural networks to detect a broad range of crops, which makes it suitable for monitoring and identifying crops at a large-scale. With accuracy benchmarks that has been accepted by government entities in Europe and Asia.
Figure AI Crop Classification for Switzerland
Grundo developed a biodiversity index based on satellite data and statistical methods into an online tool that can help tackle several use cases such as protecting forest reserves, detecting land encroachments, habitat rehabilitation, and track urban centre development.
Soil Fertility Index
Grundo’s model can ingest satellite data of a target region and estimate the content of organic matter. The model selects relevant wavelengths from the satellite data and performs calculations to quantify the amount of organic matter based on the latest research. Since organic matter and fertility are strongly correlated, we can assess the level of fertility of the soil.
Figure Our Fertility Index applied to two different regions in Indonesia. On the left, we have the region of Kalimantan, where peatlands are present, and potentially the index could be used to identify them (between brown and pink regions). On the right, we have a region with farms where the fertility index could be used for identifying portions of land that are good for agriculture. Currently, the index is based on firm theoretical models and subsequent calibration and ground validation will enhance its accuracy.
Increasing public awareness of the consequences of pesticide and fertilizer use for the environment has put great reputational and regulatory pressure on many organizations. Environmental, government and social responsibility are no longer limited discussions in academic circles and NGOs, companies are expected to disclose and monitor their activities that impact the broader natural and societal environment. Grundo is committed to the cause and has built novel and cost-efficient ways, using remote sensing, to detect chemicals and other hazardous substances that could damage the environment. Our solutions can assess the environmental impact by leveraging high-tech satellite image processing and the latest scientific research.
Figure Our methods are capable of detecting pesticide and fertilizer use with the help of satellite image processing by combining both low- and high-resolution imagery and using higher-resolution approximation algorithms when punctually required.
Soil Moisture Index
Grundo developed a model to determine soil moisture by using satellite data. The model is based on the so-called trapezoidal method and is enhanced by applying superior mathematics, allowing for more analytics insights than what's commonly available in the market.
Figure Our in-house Moisture Content Index applied to Central Kalimantan. On the left, we see the index during the rainy season and, on the right, during the dry season. As we can see, on average the index shows a good measure of the moisture content of the soil, and potentially, it can also be used to determine when a portion of peatland has become too dry increasing the risks of catching fire. Currently, the index is based on firm theoretical models and subsequent calibration and ground validation will enhance its accuracy.
Ground Stability Monitoring
Ground stability analysis and sinking of infrastructure are highly location-dependent owing to the underlying geology. With Grundo’s ability to execute on a global scale, ground stability analysis can be quantified using daily radar satellite data and can provide a time-series of an individual building at high-resolution, which can be extended up to large city-scale depending on requirements.
Figure A low-resolution ground stability analysis example for the city of Geneva, Switzerland
Maps Web Portal
Grundo Maps Web Portal is the culmination of several systems into a single web accessible interface that provides web users direct access to NDVI, Soil Fertility, Crop Distribution map overlays in near real-time. Grundo MWP is also a tool for capturing crop and geo-specific data for custom applications. The system is capable of ingesting unique crop samples that feed into Hamilton AI as training data and subsequent use.
Figure Grundo Maps Portal allows users to apply various analysis through geo-fencing specific areas, in this example, a crop classification render is imposed on the selected area.