The National Agriculture and Food Research Organization (NARO) has developed a technology that uses drone-acquired aerial imagery and artificial intelligence (AI) to easily estimate legume forage coverage in grasslands where grass and legume forages are grown together in mixed swards. Building on AI-based vegetation assessment methods previously developed for small-scale experimental fields, the technology was further refined for application in large-scale production fields. As a result, it enables highly accurate mapping of legume distribution across hectare-scale grasslands, a typical pasture size in Hokkaido. This technology is expected to support efficient and sustainable grassland management practices, including precision fertilization and overseeding, to maintain an optimal balance between grass and legume forages across entire pastures.
Overview
Grasslands used for grazing or forage production are commonly established as grass-legume mixed swards to improve forage productivity and quality. To maximize these benefits, it is important to monitor the proportion of legumes in the sward and maintain their contribution to forage at an appropriate level, typically around 30%.
NARO has previously developed an AI-based vegetation assessment method that analyzes drone-acquired aerial imagery to easily estimate legume forage coverage in images. This technology was originally developed for small-scale experimental fields used in forage cultivar evaluation (joint research with Bandai Namco Research Inc.; NARO press release, 2022: https://www.naro.go.jp/publicity_report/
In the present study, we enhanced the method by georeferencing aerial images with high-precision positional information and by enabling the estimation results from individual images to be integrated and mapped across entire grasslands. These improvements allow the technology developed for small experimental plots to be applied to hectare-scale production fields. The resulting system enables easy assessment of legume distribution over large pasture areas and is expected to support precise and efficient grassland management practices, such as targeted fertilization and overseeding in areas with low legume coverage and consequently reduced nitrogen input from symbiotic nitrogen fixation by rhizobia.
Appropriate grassland management supported by this technology is expected to contribute to greater efficiency and sustainability in dairy and livestock production as well as in on-farm forage production. By supporting stable forage production, the technology is also expected to serve as an important foundation technology for the sustainable production and supply of dairy products, thereby helping support people's daily lives.
Related Information
Funding: Operating grants and JSPS KAKENHI Grant-in-Aid for Scientific Research (C): JP21K05551 and JP23K05197
Patent: Japanese Patent Application Publication No. 2023-100107 (JP 2023-100107)
