The Rural Development Administration (RDA) has developed an automated crop trait survey technology by integrating phenotyping with artificial intelligence (AI) to accelerate the breeding of climate-resilient varieties. Traditionally, researchers manually measured crop size, shape, color, and yield, a process that was time-consuming and labor-intensive, often resulting in inconsistent outcomes due to subjective judgment. With climate change rapidly altering cultivation environments, the need for repeated and large-scale surveys has increased, highlighting the limitations of manual methods. In response, the RDA initiated the development of this advanced automation technology.
Over the past three years, the RDA utilized phenotyping and AI learning technologies to build a database of more than 3.4 million data points and completed six automated trait survey technologies. These technologies cover mushroom yield, soybean leaf shape classification, soybean growth prediction, strawberry shape classification, apple quality, and corn ear height. The automation system boasts an accuracy rate exceeding 90%, significantly outperforming traditional manual approaches. By analyzing data captured with standard RGB and hyperspectral cameras, the system quantifies and assesses crop size, shape, disease occurrence, and growth status.
Implementing this technology reduces the trait survey time from up to a week to an average of just 30 minutes. Labor requirements are greatly minimized, and the objectivity and reproducibility of results are enhanced by eliminating subjective human judgment. The RDA has filed four patents, published one research paper, and registered one copyright related to these technologies, solidifying their research achievements. Furthermore, the RDA is actively disseminating the technology through training sessions for domestic researchers and transferring it to industry partners.
The AI-powered phenotyping automation technology received the Encouragement Award at the ‘2025 Public AI Transformation Challenge’ hosted by the Ministry of the Interior and Safety, recognizing its potential and impact in the public sector. Kwon Su-jin, head of the Digital Breeding Support Division at RDA, emphasized that this technology will play a pivotal role in data- and knowledge-driven agricultural research. The RDA is committed to strengthening the digital breeding foundation and expanding the adoption of field-oriented technologies. This innovation is expected to accelerate the digital transformation of agricultural research and significantly enhance the efficiency and accuracy of new variety development.
The RDA’s AI-powered crop trait survey automation marks a significant leap in agricultural digitalization. By leveraging big data and AI, the technology enables faster and more accurate breeding, allowing for agile responses to environmental changes such as climate variability. As this technology spreads across research and industry, it promises to optimize labor, boost productivity, and reinforce data-driven scientific decision-making in agriculture.