Viridis RS

There are 500 million smallholder farm households in the world, who have less than two hectares of land to farm. Increasing risks of crop failure due to climate change and growing population makes it more urgent to increase the productivity of these farmers. 

Lack of information on what these farmers are growing and their yield makes it difficult to help the farmers to do so. Aerial survey of the farms can help by collecting data over large number of farms. 

We propose using light manned aircraft such as Cessna or microlight, which can fly higher and faster than drones, making it more cost effective and scalable, and can fly under the cloud unlike satellites. 

We intend to use the data collected to provide individualized actionable information to the farmers to prevent crop loss while enabling insurance companies and banks to provide services to these farmers. 

We are looking to provide better data to both smallholder farmers and the companies that are looking to service the farmers. Smallholder farmers' productivity has stalled or declined in the recent years and increasingly unreliable weather due to climate change further put their livelihood at risk. 

Farmers need credit to invest in productivity enhancement such as fertilizer and better seeds but lack access to credit due to lack of data. Most farmers do not have access to crop insurance since collecting data on what the farmer plants and measuring yield is very expensive if done traditionally. Pests and disease destroy approximately 20-40% of crop. Farmers lack information on how to deal with them effectively. Pests and disease are migrating into different regions due to climate change, exposing farmers to unfamiliar problems and making this problem more urgent to address. 

Using specialized cameras, it is well known that we can tell health of plants. We fly these cameras on light aircraft to monitor farms of smallholder farmers who typically has less than one hectare of land. Airplane can fly higher and faster than drones, making it cheaper to monitor large area. It can also fly under the cloud and provide higher resolution than satellites, which struggle with small plot sizes. 

The data collected can be analyzed to detect pests and disease before it becomes visible to the naked eyes, and provide data like type of crop grown and estimated yield to insurance companies and banks so that they can insure and lend to the farmers. 

Multispectral camera data is collected on regular basis to provide timely advice and improve the accuracy of the yield estimate. Machine learning is used to analyze the data for actionable information. Data collection process has been optimized for minimum cost and maximum insight which is a product of PhD research at MIT by Dr. Mark Jeunnette, co-founder of Viridis RS. 

Our solution is targeted towards the 500 million smallholder farmers around the world, many of whom are subject to greater uncertainties due to climate change. Warming climate makes rainfall more irregular and change the patterns of pests and disease, increasing the likelihood of crop failure for these farmers. Regular monitoring can help provide them with early warnings on pests and disease along with information on how to deal with the infestation. Greater transparency and data on these farmers can help insurance companies and banks provide badly needed services like crop insurance and loan to improve productivity. 

We are working with Deshpande Foundation in India and engaging with farmer producer organizations (FPOs) to understand their needs and collect information. Cash crop farmers such as cotton farmers are willing to pay for actionable data that can reduce their crop loss to pests. 

Our data will enable us to provide actionable information to the farmers to prevent crop loss via e-extension services and local FPOs. It will also enable insurance companies and banks to provide much needed crop insurance and credit to the farmers so that the farmers' risk is reduced and they can make productivity improving investment. 

We aim to support the small scale farmers through better information and enable other companies to provide them with superior services by providing them with necessary data. We expect to improve productivity through reducing loss by pests and disease, and more investment through helping to enable better access to insurance and credit. 

Also understanding where and how much crop is grown is key to help crop aggregators to engage smallholder farmers more effectively and improve supply chain. We hope to engage crop aggregators, fertilizer and pesticide producers as well as companies that contract farmers to help them improve their operation. 

Most of our competitors are using satellite data or drone. Satellites cannot see through the cloud. Since the major crop season in many parts of the world coincide with rainy season, it is of limited use other than rice which can be detected with synthetic aperture radar. Drones have limited range and are not allowed to fly above 400ft in most of the world, making it expensive and difficult to scale. Light aircraft such as Cessna 172 or microlight are inexpensive and can take imagery over much larger area. Using our algorithm which minimizes the flight time required, one airplane can cover 40,000 hectares vs. a drone which can at best cover 1,000 hectares per day. 

In addition, instead of creating a map, we process the data to preserve the oblique angle vision as well as orthogonal vision, increasing the accuracy and the insights. 

Our company was started to put into real life application the research by Dr. Mark Jeunnette. Talking to farmers in India, Mark identified the problem that while several extension services are available to the farmers, the extension services did not know what was happening on the farm therefore could not give actionable advice to the farmers.

Any data service trying to solve this problem has to be cheap and scalable. This meant that we should leverage existing technology for better output. Mounting camera that is typically used with a drone on an airplane gives the same information but is much cheaper and can cover larger area. Mark's algorithm minimizes the cost and maximize the information gathered from the survey.

By collecting data at plot level and on regular intervals, we expect to be able to outperform competitors in accuracy for determining crop type, health and yield based on imagery. 

Using remote sensing for assessing crop is widely used for commodity traders, insurance and precision agriculture service companies. In the US, drone and airplane are used as well as satellite to provide information to the farmers. In India, satellite data is used by crop insurance companies and banks. However because of the small size of the plots, the data has been of limited use despite many companies looking into this problem. Drones are used in small scales hampered by high cost. 

Dr. Mark Jeunnette's paper link: https://www.spiedigitallibrary...

For a long time, small scale farmers have been mostly ignored or taken advantage of. There is low level of trust in government or organizations. In turn, the companies have low level of trust in the farmers due to lack of data. 

We believe that by providing individual level advice on pests and disease to the farmers that they benefit from, we can earn their trust in our data. In the short term this will enable the farmers to reduce their crop loss to pests and disease. 

Once trust is built on our data, it can then be used to build financial products around. For example, the data can be used to assess crop insurance payout. This will only work if there is trust in the data from both parties. 

If we can build independent way of surveying large number of farms and assess the crop grown, health and yield, this enables companies to serve the farmers better. For example: 

Banks can assess credit worthiness of the farmer by understanding historical yield. Also the banks can control their risk by monitoring for possible large scale crop failure. Better access to credit is crucial for farmers to improve their productivity. 

Insurance companies can provide better insurance product that pays based on the yield of the farmer rather than proxy measure such as weather. The low take up rate of weather insurance in developing countries is due to the gap between what weather station records as rainfall which is often very far from the farm, and the yield achieved by the farmers. Crop insurance has been shown to increase farmers' investment into productivity enhancing measures.

Lack of understanding of which farmer is growing what crop hinders government from providing help with input such as fertilizer and limit the markets the farmers can access to sell their crop. Data on amount and location of crop grown for each type will help companies build better supply chain and give farmers access to more markets helping to increase the profitability.

We are planning a pilot project which will cover approx. 5,000 hectares of farm and expect to work with 1,000 farmers. This project is to test the accuracy of the model. We are planning two more projects in Uganda and Sri Lanka. We hope to secure enough funding and paying customers to serve at least 10,000 farmers in one year. In five years if we can get significant traction particularly in India we expect to serve at least 500 thousand farmers. Crop insurance in India is awarded to insurance companies at a district level. Each district typically has over 1 million people. Getting commission for three or four districts will let us cover at least half a million farm households. 

Our goal is to first reduce the loss of crop due to pests and disease, and help the farmers improve their productivity. 

Secondly we aim to improve access to services such as insurance and credit through providing better information to the companies. We aim to work directly with insurance companies and banks to improve the products provided to the farmers. Our team has knowledge and reach into the insurance industry to scale the solution. We have also discussed our solution with World Bank and several rural microfinance organizations in Africa and SE Asia who are interested in implementing the solution. We plan to work with microfinance companies, banks and insurance companies to expand to other countries. 

Regulation is a big barrier as some countries have laws limiting use of aerial survey. India has one of the strictest rules which has slowed our progress. 

Data collection to demonstrate the product is more costly than competitors since although airplane is cheaper than drone or high resolution satellite at scale, the minimum cost is high.

Agriculture is a low margin business and therefore we will need to demonstrate significant advantage over services that use free satellite data. 

Regulation: We have built deep knowledge of Indian regulatory environment and are making progress on obtaining permission. We have also made contact with potential aviation partners in several countries who can help us navigate the regulatory environment. 

Cost of data collection and demonstration of significant advantage: We are investing equity funding we have raised for our first pilot project for demonstration 

We are also working with several NGOs and farmer organizations who work with farmers to gather ground truth data, which is a very important component to developing optimal model for crop assessment. 

Two full time staff, one part time. We are working closely with another startup which has 11 full time staff.

Heejae Cho: Heejae is a long time (re)insurance underwriter and actuary. Heejae met Mark during her MBA at MIT and decided to start Viridis RS with him to solve the problem of inadequate crop insurance in the developing countries. Heejae's insurance experience enables her to understand the data problem from the clients' perspective and path to scale the business. 

Mark Jeunnette: Mark has PhD in Mechanical Engineering from MIT and his research forms basis of Viridis RS. Mark has worked for two years in Ethiopia developing engineering solutions for smallholder farmers with International Development Enterprises as well as at IDEO and BMW. Mark has deep understanding of remote sensing technology and how it applies to crop sciences.

Aditya Bhat: Aditya is CEO and cofounder of AirProbe, a computer vision analytics company that analyzes drone imagery for solar panel inspection. Aditya aims to use technology to reduce the impact of climate change which made him become interested in agriculture as the next area to focus. Aditya has B.Tech in Electronics and communication and one of the pioneers in AI driven solution for solar panel inspection. 

Our team marries the scientific rigor and technical skills for remote sensing and AI / machine learning for data analysis as well as deep knowledge of business. 

We are working with Deshpande Foundation, a non profit organization that works with farmers in India. Deshpande Foundation will help us collect the data from the farmers which is crucial for developing accurate model. 

We collect the aerial data of the farmers for companies that are interested in that data. In return, we provide pests and disease warning to the farmers at low / no costs through local extension services.

Banks, insurance companies and any other companies that work with smallholder farmers need better information to provide necessary services and products for them.