PlantumAI utilizes Artificial Intelligence technology to augment farmers in less-fortunate areas, providing them with tools to identify/treat crop diseases, advising them on optimal planting and harvesting times in a rapidly-changing environment, and linking them to local universities for communication.

For the last two years that PlantumAI has been in place, more than 1500 crop disease instances have been detected. PlantumAI assists 23 villages in India. Farmers can more reliably provide crops for their areas, while also reducing their use of toxic pesticides.

The algorithms for crop disease detection are custom Neural Networks (NNs), developed through PyTorch and structurally optimized through ENAS (Efficient Neural Architecture Search). The Convolutional NN, to detect crop disease from smartphone images, learned critical features from a large dataset of plant pathology images collected through both online datasets, and a curated database in partnership with Akola Agriculture University. The app was developed in Java, as the majority of farmers have Android-based phones. The server-backend is in Python, dealing with distributing data between the farmers and the university, allowing the university to track outbreaks of crop disease in specific areas, enabling them to provide directed education about the crop disease. The crop planting/harvesting times advice algorithm is a Dense NN on the backend, as environmental values are pulled to update the algorithm for individualized users, in real-time.

PlantumAI utilizes novel technology to optimize the planting/harvesting methods in underserved areas through the efficient prediction, detection, and prevention of crop disease, all with a practical and proved implementation.

What inspired you (or your team)?
I visit India, every summer, to meet up with family. It never registered, for me, how there were many things that I took for granted, living in the US, that my family in India didn’t have the privilege of having.

During my visit to India in 2016, when I visited a farm that my grandpa worked on, he made me aware of the medical and environmental problems that over-pesticide use was causing. Many people were falling seriously ill because farmers not only used a single pesticide, but instead, showered their crops with many in the hope to eliminate all chances for disease. The pesticide leaked everywhere, including into the water supply of many people.

I was horrified that this issue was not being resolved, as many innocent people’s health and safety were in grave danger. Initially, I was upset at the farmers for not regulating the pesticide use, but I realized that they too were under pressure to harvest enough food to feed their family and their region. If a pathogen destroyed their crop, then there would not be enough food to feed everyone, which was another dire scenario.

Crop disease is a significant problem because an infection can completely devastate the harvest of a farmer, significantly reducing their profits, along with reducing the amount of food for their area. In this regard, farmers are under constant pressure to achieve a plentiful harvest every year, which results in them using a large number of toxic pesticides whenever disease is noticed on their crops. Unfortunately, many of them do not have the exact knowledge of crop disease, so to completely eradicate the disease, they use a multitude of pesticides to leave nothing to chance. Unfortunately, this is not safe, as the pesticides are poisonous, and this process is not cost effective.

Another big problem is the problem of climate change, which is significantly affecting the techniques and procedures that have been standard for many years. The change in environmental conditions is now more unpredictable than ever.

My desire was sparked to create a solution. I spent the next year researching, developing, and prototyping. In the summer of 2017, I returned to India, to test my prototype, the first iteration of the PlantumAI app.

PlantumAI was created to make both a lifestyle and safety change within less-fortunate populations, to allow them access to technology that enables them to save their food supply, without resorting to toxic contamination of their environment.