식량 작물의 확보 및 생산량 예측은 국가 발전에 있어 필수적이며, 국가 경제뿐만 아니라 전 세계 식량 안보에 기여 한다. 최근 환경오염으로 인한 이상기후는 식량 작물 생산량에 직ㆍ간접적으로 부정적 영향을 끼치고 있어, 작물 수확량 예측 불확실성이 높아지고 있다. 특히, 노지 작물의 경우 생산량 감소와 품질 저하 문제가 화두 되고 있다. 이러한 문제는 농가들뿐만 아니라 소비자들에게도 큰 피해를 안겨주고 있다. 이러한 생산량 예측 이슈를 해결하기 위해 최근에는 인공지능 기술이 농업 분야에도 활발히 적용되고 있다. 작물 수확량의 정확한 예측을 위한 머신러닝 기반 연구가 집중적으로 수행되고 있다. 따라서, 본 연구에서는 이와 같은 인공지능 기반의 노지 작물 수확량 예측 기술(머신러닝, 딥러닝, 하이브리드 모델 등) 현황 및 작물 수확량에 가장 영향을 많이 끼치는 모델 파라미터 등을 조사하였다.

This study was conducted to provide basic data for high-throughput screening (HTS) system construction based on phenomics. Rice (Oryza sativa cv. Chucheongbyeo) seedlings in vegetative growth stage were grown in the glass house and treated with 0, 3.75, 7.5, 15, and 30% (w/v) of polyethylene glycol (PEG) to give osmotic stress. Three days after PEG treatment, hyper-spectral reflectance images were obtained and analyzed after removing background image in several steps. The reflectance of rice seedlings treated with 15 and 30% of PEG solutions were significantly different at 680 nm, where differences in the chlorophyll reflectance spectrum and visual symptoms were not observed. These results thus indicate that hyper-spectral reflectance observed at 680 nm can be used to screen drought tolerant rice lines. A HTS system equipped with this hyper-spectral reflectance system may play an important role of future rice breeding program.

This study was conducted to investigate plant body temperature response of soybean (Glycine max) to saline stress. Two-weeks-old seedlings of soybean in V1 growth stage were treated with 0, 10, 20, 40, 80 and 160 mM of NaCl for salt stress. Thermal images acquired using Flir T-420 (US) were obtained at 4 days after treatment. Soybean leaf temperature increased with increasing NaCl concentration, resulting in significant positive correlation between soybean leaf temperature and stress intensity (P < 0.01). Leaf temperature of soybean was significantly different at 160 mM of NaCl, where no visual symptom was observed. Therefore, soybean leaf temperature can be used for evaluating the response of soybean to salt stress as a non-destructive and phenomic parameter. Non-destructive diagnosis of soybean leaf temperature may be a key parameter in a high throughput screening (HTS) system in breeding program for salt stress tolerance soybean cultivars.