Abstract
Vegetable crop productivity is susceptible to damage from weed competition, with early season weeds a control priority to prevent significant yield loss. There is an urgent need for a reliable robotic sensing system that can work well in a variety of crops to achieve universal weed/crop differentiation, which would facilitate further development in robotic technologies for farming and bring economic benefits to vegetable production. The aim of this study was to develop a novel technique to create a machine-readable crop plant using a systemic crop signalling compound applied to seeds or transplants. The protocols for the crop signalling method and its detection are described. Rhodamine B (Rh–B) was selected as the signalling compound in this study, because it could be used as a fluorescent tracer, had a unique optical appearance in plants, and had the necessary properties to allow systemic behaviour in vegetable seedlings. The Rh–B tracer was applied to snap bean and the systemic behaviour analysed with a fluorescent macroscope. The uptake of Rh–B varied among treatment methods. The Rh–B uptake through the seed coat of snap beans was found mainly in the seedling hypocotyls. The results for root uptake showed that Rh–B could be more readily transported to the whole plant through the root system as compared to the application to seeds. The midvein and secondary veins of bean leaves showed stronger Rh–B fluorescence than other regions of the leaf. Higher concentrations of Rh–B resulted in greater absorption by the plant. Although the crop signalling compound could follow both seed and root pathways for plant uptake, the uptake based on the root pathway had greater capacity than that of the seed pathway. The use of Rh–B provided a systemic crop signalling compound was discussed on further research and field tested to have application to enhance weed/crop differentiation by automated weeders in vegetable crops. The systemic crop signalling system successfully created a machine-readable signal on vegetable crops, and appeared to be non-destructive, cost effective, efficient and accurate for performing automatic plant care tasks.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 62-74 |
| Number of pages | 13 |
| Journal | Biosystems Engineering |
| Volume | 193 |
| DOIs | |
| State | Published - May 2020 |
Bibliographical note
Funding Information:The authors would like to acknowledge the funding from the United States Department of Agriculture , National Institute of Food and Agriculture , under the Specialty Crop Research Initiative (SCRI) grant ID USDA-NIFA-SCRI-004530 , the California Tomato Research Institute, and the California Leafy Greens Research Program. The authors also would like to acknowledge professor Alan G. Taylor in Seed Science and Technology, from Cornell University, Leland Neilson and John S Rachuy from the University of California, Davis for their technical assistance in the completion of this research.
Publisher Copyright:
© 2020 IAgrE
Keywords
- Automated weeding
- Bandpass filters
- Fluorescence macroscope
- Fluorescent compound
- Green light excitation
- Robot-plant interaction
