Non-invasive plant phenotyping

The practical relevance of plant phenotyping is its high throughput, when applied on a larger scale (e.g., field), without significant demand for resources (time and finances). For that purpose, we aim at developing a lower-cost, non-invasive, image-based plant phenotyping system that can be mounted on unmanned aerial vehicles (UAVs or drones). Both hardware and software will cost significantly less than existing equipment, while retaining performance, and reducing financial investments. Our idea relies on the data modeling of the gap between the performances of high- and lower- end equipment, by which AI-assisted algorithms using lower-end (multispectral and vision) cameras will utilize the knowledge acquired by algorithms using high-end multispectral imaging equipment.