Folia: Utilizing Disease Detection in the Agriculture Industry
Exploring Folia, a Convolutional Neural Network (CNN) that uses feature extraction to pinpoint common foliar diseases.
The Problem
According to the U.S. Department of Agriculture, every year, up to 40% of global crop production is lost to pests and the diseases they carry. As such, plant disease costs the global economy around $220 billion dollars annually. In a world where modern agricultural practices are already exacerbating cropland and its resources, such a loss should be prevented wherever possible.
Our Goals
- Devise a strategy for leveraging technology to mitigate crop loss and reduce the prevalence of foliar diseases within agricultural contexts.
- Implement this strategy in a manner that effectively mitigates the perennial health hazards inherent to prolonged outdoor agricultural engagement, encompassing risks such as heat-related ailments.
- Execute these endeavors with a focus on economical viability and seamless expansibility, ensuring the feasibility of widespread adoption and integration.
The Solution
What is it?
Folia is a Python-based Convolutional Neural Network (CNN) loaded onto a Raspberry Pi camera (held together by a 3D-printed Polylactic Acid chassis) that utilizes feature extraction to detect the beginnings of common rust, gray leaf spot, and northern leaf blight on plant leaves.
Once detected, the system sends push notifications to a designated user’s phone, alerting them of the type of disease and an approximate location (via coordinates).
Why do this?
Aside from the obvious financial loss that crop loss poses to the mere 2% of farm owners in the U.S. population, I believe that as the world moves towards a future of AI-based technologies, the agricultural industry needs to do the same in order to minimize the endangerment of human lives. After all, according to the CDC, farm workers are 20 times more likely to die of illnesses related to heat stress. It’s also important to note that as the younger generation becomes more and more integrated into tech spaces, it would be beneficial to solve industry issues in ways that would be most familiar to them, right from their cell phones.
Once implemented, Folia is expected to be able to save the lives of young farmers by allowing an automated system to do some of the work for them, thereby limiting their overall outdoor exposure.
Next Steps
Though the algorithm is not only effective but also accurate as is, there’s always room for improvement. My projected next steps include:
- Expanding the feature extraction model to be able to accommodate more foliar diseases, rather than just the most commonly occurring ones.
- Test the effectiveness of Folia in more remote regions that may have less network stability.
- Present Folia at the 2024 UN Science Summit.
Partners
I’m grateful to have received over $10,000 USD in funding for this project from the organizations below. Their support is much appreciated.
