Team #11 // Budnip
Crop diseases and pests persist to be a major problem around the globe, causing huge losses to farmers, threatening food security and damaging agriculture/trade-dependent economies. It is estimated that crop pests and diseases account for up to 68% of the world-wide production tonnage loss in the 21st century (Oerke, 2006). Even though plant diseases and pests are inevitable, the key to minimizing the damage relies on the ability to identify and address them before they reach epidemic proportions. Our solution aims at spotting new outbreaks based on satellite technology, that allows to detect them from an early stage, cutting the losses significantly. After getting sufficient data we plan of offering macroscopic crop mapping services to industries related to agriculture in the long term.
Current techniques used to track crop development and detect plant diseases and pests involve a huge degree of manual testing and, even the most advanced techniques, are limited to a field level study. This conveys that monitoring farms becomes very expensive and that the miscommunication between farmers make it hard to properly detect crop diseases. By the time a proper comprehension of the situation is reached, it is often too late.
Target user / Customer
Governments would be the initial customers and users of our solution.The losses due to crop epidemics are huge, so governments would have an immediate incentive to invest in disaster prevention measures like our solution. As we grow, we will accumulate more agri-data, which would enable us to provide more features like macroscopic crop scanning to predict harvest, future crop prices etc, which will give our services a new market in insurance companies and agricultural product consuming companies(eg Pepsi, Mondelez etc). Thus in the long term our plan is to diversify our business, using the short term goal of preventing epidemics as a launching pad
Our solution and how the concept is feasible
Our solution aims at an early anticipation of crop disease and pest outbreaks to prevent them from spreading further, cutting losses of the crop yield. Budnip uses satellite data from Sentinel-2 to monitor crop fields with a frequency of 5 days. In-built bands such as VIS-NIR, fluorescence and thermal sensors, SAR and Lidar systems can be used to map biophysical variables such as leaf chlorophyll content, leaf water content and area index, which are good indicators of the health of the crops. By monitoring their evolution after time, our algorithm is capable of detecting anomalies, which might be caused by pest and plant disease outbreaks.
Figure 1. The image on the left shows the non-filtered image of Italian crop fields, whereas an NDVI filter is applied on the right one, which maps healthy vegetation into light colours and unhealthy into darker ones.
The customer is immediately informed about any detected anomalies, who may take appropriate action. In the future, our goal would be to offer a more complete solution by also being able to identify the diseases and pests behind these anomalies. This would be done by training machine learning models that learn from previous disease outbreaks and the circumstances that surround them.
Four unique value propositions
- Disease and epidemic detection at an early stage. Images of the same spot on earth are received every 5 days, due to the orbiting time of the satellite, which is sufficient given the disease spreading rate, and much faster than current techniques, which require many time-consuming intermediate steps.
- Centralized system. Our solution aims to be the first one to achieve holistic monitoring of the crop fields and implement a centralized system, reducing costs and correlating multiple land data points.
- High accuracy. We have access to plenty of data from all around the globe, which allows us to train machine learning algorithms that will make more accurate predictions. This accuracy will only increase as we gather more data from the newer satellites that will be launched
- Low-cost. Little infrastructure is required, since it uses already existing satellite technology, as opposed to current solutions that depend on manual labor or drones, which is often inaccessible for small farmers.
The business model of our solution would be establishing contracts with the customer(initially governments) to provide a continuous service that informs the customer of the evolution of the crop fields and warns them if any anomalies were detected.
Short term: only focus on organizations at a regional/ national level. As a side business for financiation during the growing face, data could also be sold to interested groups such as insurance and trading companies(B2B).
Long term: scale to global monitoring, with international customers and develop the capability of diagnosing diseases and pests, to expand further in the market.
Crop losses will be significantly reduced, helping to combat food shortage and world hunger. (SDG 2)
Contribute to the stability of developing countries, where agriculture is crucial. (SDG 8)
Minimize collateral damage to trading nations and farmers.
We are a very diverse team, representing 3 countries. Blanca is from Spain, and specializes in Cyber systems and has experience in ML, data analysis and project management. Adithya is from India, and has previous experience in GIS, satellite spectroscopy and Deep learning on images. Julian is from Denmark and has local knowledge about business development and general engineering. We are motivated by the fact that there are immediate solutions we can provide by existing satellite infrastructure that have large scale impacts on food security.
List full name and e-mails (preferably not your student e-mail)