At the Advanced Radiation Research Institute of the Korea Atomic Energy Research Institute located in Jeongeup-si, Jeollabuk-do. Upon visiting on the 17th, entering a glass greenhouse spanning 660m² (approximately 200 pyeong), over 200 pots of lettuce were seen moving in line on a conveyor belt. The size and shape of the lettuce varied. Some reached a height of 1m, while others were only 20cm. The pots, moving in line, entered a tunnel on the right side of the glass greenhouse and stayed inside for about 3 minutes before being automatically released. Kim Sang-hoon, head of the Radiation Breeding Research Laboratory, explained, “The lettuce pots enter the tunnel once or twice daily to be precisely inspected by state-of-the-art cameras to determine if they are superior 'radiation breeding' varieties.”

● Hyperspectral Cameras Deployed to Identify Superior Varieties

Plant radiation breeding is a technology that creates new traits rarely seen in nature by irradiating embryos, seeds, and seeds. As the climate crisis threatens human survival, developing national strategic crops that thrive in extreme climates is urgent. Radiation breeding is a simple development process and does not involve gene modification or manipulation, so radiation-bred crops are already actively sold in markets in Europe, Japan, and India.

On this day, the appearance of over 200 lettuces in the glass greenhouse was all slightly different. Surprisingly, the seeds of all the lettuce were planted in the pots on the same day, August 18. Even the same lettuce can produce various varieties depending on the speed, time, and direction of radiation exposure. The more precisely examined, the more different the appearance becomes.

The lettuce undergoes precise appearance inspections daily in the tunnel within the glass greenhouse. The tunnel is equipped with an 'RGB camera' that captures visible light, a 'fluorescent camera' that observes chlorophyll fluorescence, and a 'hyperspectral camera' that observes high-wavelength light. Cameras of various wavelengths observe plants daily, precisely analyzing what characteristics the varieties possess.

The Advanced Radiation Research Institute, equipped with such facilities, is a core hub for domestic radiation breeding research. Since 2018, it has been enhancing technology based on 'Phenomics.' Phenomics is a biological field that studies how the proteins, metabolites, and genetic components of organisms determine the appearance or function of plants.

Kim stated, “It is the concept of analyzing the appearance of plants to understand their characteristics,” adding, “Recently, with the advent of artificial intelligence (AI) technology, the technology for selecting radiation breeding varieties is advancing.” In the past, the process of manually selecting varieties by human observation over several years can now be processed in a short period using automated sensors and AI.

● Selection Accuracy Enhanced with AI Technology

The fluorescent camera inside the tunnel measures the photosynthesis efficiency of plants, while the hyperspectral camera measures the composition and metabolic changes of plant components. During photosynthesis, when chlorophyll absorbs light, some energy is converted into chemical energy, but the rest is emitted as fluorescence. The fluorescent camera captures this process to assess photosynthesis efficiency. Lower fluorescence indicates better photosynthesis.

The hyperspectral camera can simultaneously measure the reflectance of hundreds or thousands of wavelengths, revealing the composition of molecules such as chlorophyll, moisture, protein, and cellulose that make up plants. Jinbaek Kim, a senior researcher at the Radiation Breeding Research Laboratory, said, “When selecting varieties through radiation breeding, it is important to analyze the specific components of the variety to understand the mechanism by which certain traits appear,” adding, “The hyperspectral camera replaces the process where researchers had to grind plants to analyze components.”

Research based on Phenomics can provide scientific evidence for why a variety was selected. Senior researcher Ko Chan-seop of the Radiation Breeding Research Laboratory explained, “Since people eat lettuce leaves, one of the main tasks of lettuce breeding is to delay the formation of flower stalks,” adding, “Through measured component data, it is possible to understand the growth rate of flower stalks before they even form.”

The Advanced Radiation Research Institute has collected data on lettuce, wheat, and peppers over eight years. The accumulated data combined with AI technology is expected to significantly enhance the accuracy and speed of variety selection. Recently, the Advanced Radiation Research Institute used 'CNN-Transformer' deep learning analysis, an image-based AI technology, to train an AI model with data observing the entire growth process of eight wheat varieties. As a result, the AI model successfully distinguished wheat varieties with 94% accuracy and wheat growth stages with 99% accuracy. CNN-Transformer is a technology that analyzes how a subject changes over time through images.

Researcher Kim stated, “By combining weather observation data with plant observation data, the speed of developing new varieties responding to climate change will significantly accelerate,” adding, “We aim to develop selection AI algorithms and intensify research to develop national strategic crops based on radiation breeding.”

Popular News