Arbora – A Digital Future for Agriculture

Much of the innovation in agriculture has focused on task execution at large scales, leaving farms with fewer resources behind, despite them being most vulnerable to the impacts of environmental and economic uncertainty. Computer Vision and Artificial Intelligence present clear opportunities to support these farmers by aggregating and interpreting complex information, simplifying problem identification and facilitating data-supported decisions. The challenge lies in bringing data-science to the farm gate by designing a user-friendly system that fits seamlessly into small-scale agricultural contexts, providing meaningful, actionable insights, and crucially, empowering farmers to focus on what they do best - farm.

Arbora is a powerful computer-vision system that accompanies farmers on their day-to-day tasks. Attaching to various farm equipment, Arbora's two integrated camera systems and onboard processor simultaneously record, map, and interpret visual information as it moves through the environment. Insights like plant performance or pest and disease identification are accessible through a computer interface. Designed to be as simple as possible to minimize implementation effort, and flexible enough to adapt to any farming context, Arbora is the first look at what a digitally-enabled farming practice might look like when considering the needs of independent and small-scale farmers.

Enabling data-informed decisions empowers farmers to develop and execute solutions while considering long-term implications, minimising risk as they transition to more sustainable and profitable farming practices. Having concrete, geographically contextualised data can help optimise crop treatment, diagnose plant or soil problems, and even facilitate the development and selection of new varietals or farming methods. Farmers can more accurately predict harvest yield and timing, making labour forecasting and offtake agreements more reliable. Even beyond the farm gate, in-depth data can reveal regional trends and aid in environmental and economic policy decisions.

Arbora is simple, flexible, and powerful, due to the intense pressures and conflicting priorities of day-to-day farming activities. The interface was radically simplified with an e-ink screen and physical buttons for all-weather use with gloved or dirty hands. The data storage module is integrated into the battery module allowing for simultaneous charging and data transfer, reducing additional processes for the user. The mount is standardised Arca-swiss, allowing the users to purchase appropriate mounts from the open market. Onboard processing mitigates the need for cloud computing, circumventing unreliable wireless networks. The three front cameras act as the navigation system alongside the onboard GPS, ensuring accurate localisation no matter the connectivity and mapping the farmer’s movements through the farm, ensuring no sectors go unmonitored. The two side cameras capture high-resolution images, allowing for yield prediction through counting buds, or monitoring pest or disease-affected plants by assessing leaf colouration. The system is algorithm agnostic, meaning farmers can select the computer-vision model that best meets their use case, and can change it depending on season, crop, or performance updates as new models are released. Arbora has been designed with open-source components for repairability and longevity, ensuring that farmers can maintain autonomy and self-reliance.