Vinobot and Vinoculer: Two Robotic Platforms for High-Throughput Field Phenotyping
In this paper, a new robotic architecture for plant phenotyping is being introduced. The architecture consists of two robotic platforms: an autonomous ground vehicle (Vinobot) and a mobile observation tower (Vinoculer). The ground vehicle collects data from individual plants, while the observation tower oversees an entire field, identifying specific plants for further inspection by the Vinobot. The advantage of this architecture is threefold: first, it allows the system to inspect large areas
... a field at any time, during the day and night, while identifying specific regions affected by biotic and/or abiotic stresses; second, it provides high-throughput plant phenotyping in the field by either comprehensive or selective acquisition of accurate and detailed data from groups or individual plants; and third, it eliminates the need for expensive and cumbersome aerial vehicles or similarly expensive and confined field platforms. As the preliminary results from our algorithms for data collection and 3D image processing, as well as the data analysis and comparison with phenotype data collected by hand demonstrate, the proposed architecture is cost effective, reliable, versatile, and extendable. identify specific areas influenced by biotic and/or abiotic stresses. By combining a new hardware architecture, new algorithms, and state-of-the-art sensory devices, these two platforms can be deployed on existing farms, at a much more reasonable cost, specially when compared to alternatives involving aerial surveillance [5, 6] or confined field platforms  . In the next section, we briefly survey platforms for phenotyping and show cost comparison between our proposed architecture and other approaches in the literature, in special those reliant on aerial vehicles. In that same section, we also summarize the technological challenges in field-based phenotyping, namely: autonomous operation, plant imaging, and crop characterization in outdoor conditions. In the last three sections, we present in detail the proposed robotic architecture, the preliminary results, and conclusions. It is important to mention here that our goal in this paper is not to investigate current or new traits for phenotyping, let alone to establish their correlation with the physiology, development, or the behavior of plants. Instead, our goal is to show that the architecture, the sensors, and the algorithms for imaging proposed here lend themselves to a reliable, accurate, and fast approach that can be successfully employed to extracting any current or new trait.