초록

Ⅲ. 연구개발의 내용 및 범위
본 연구에서는 종자의 발아 가능성 예측을 위한 종자의 분광특성을 구명하였으며, 분광정보와 영상정보를 동시에 측정 가능한 초분광 형광 영상처리 기술(400 ~ 700nm)과 초분광 반사광 영상처리 기술(400 ~ 1000 nm)을 이용하여 발아 예측 기술을 개발하였다. 또한 측정부의 크기 및 형광의 영향 등 기존 라만분광법의 한계를 극복할 수 있는 광대역 라만 분광법을 이용하여 발아예측 기술을 개발하였다. 개발한 발아 예측기술을 적용하여 우량종자 선별 기술 개발에 관한 연구를 수행하다.

Abstract ▼

Current germination evaluation is a destructive, labor-intensive, time-consuming, and expensive method. Thus rapid and non-destructive assessment of seed viability is required in the seed industry. In this study, the sorting system for classifying of viable or non-viable seed was developed using hyp

Current germination evaluation is a destructive, labor-intensive, time-consuming, and expensive method. Thus rapid and non-destructive assessment of seed viability is required in the seed industry. In this study, the sorting system for classifying of viable or non-viable seed was developed using hyperspectral imaging techniques and Raman spectroscopy. Hyperspectral imaging techniques have advantages of spectroscopy and imaging.
Partial least square discrimination analysis(PLS-DA) models for classification of viable or non-viable seed were developed using hyperspectral fluorescence imaging with the range from 420 to 700 nm with UV-A excitation and hyperspectral reflectance imaging with the wavelength range from 400 to 1000 nm. The results of developed PLS-DA models using the fluorescence spectra showed that the viability discrimination accuracy of cucumber, pumpkin, and pepper seed were 96.7%, 95%, and 95%, respectively. The discrimination accuracy of PLS-DA models using the reflectance spectra of pepper, pumpkin, and cucumber seed were 100%, 98.3%, and 96.7%, respectively. These results showed that hyperspectral imaging technology has good potential for discriminating nonviable seeds.
The PLS-DA models to predict viable seeds were developed with measured transmission and back-scattering spectra from Raman spectroscopy in the range from 150 cm-1 to 1890 cm-1 and germination test results. Raman peaks in The Raman spectra have features of polyunsaturated fatty acids. Various types of spectra pretreatment were investigated to develop the classification models. The results of PLS-DA models developed by the transmission spectra of cucumber, pumpkin, and pepper seed showed 100%, 98.3%, 96,7% discrimination accuracy, respectively. The discrimination accuracy results of PLS-DA models using the back-scattering spectra of cucumber, and pepper seed were 100% and 96.7%, respectively. These results showed that Raman spectroscopy technologies can be used to select the high quality seeds.
A nondestructive on-line seed sorting system was developed using a hyperspectral imaging system with the range from 400 to 1000 nm. The system was constructed with a low-light sensitive electron-multiplying charge-coupled device(EMCCD) camera, an imaging spectrograph with C-mount lens, a reference measuring unit for calibration, and a sample transfer unit which conveying the seed during measurements. The results classified the viable cucumber seed using this system showed 99% accuracy. This results indicate the developed seed sorting system can potentially be used as a rapid and non-destructive method for discriminating viable seeds.

목차 Contents

• 표지 ... 1
• 제출문 ... 2
• 요약문 ... 3
• SUMMARY ... 5
• 목차 ... 7
• 제 1 장 서 론 ... 8
• 제 2 장 국내외 기술개발 현황 ... 9
• 제1절 국내 연구 현황 ... 9
• 제2절 국외 연구 현황 ... 9
• 제3절 국내외 연구 현황 비교 및 필요 연구 분야 ... 10
• 제 3 장 연구개발수행 내용 및 결과 ... 11
• 제1절 종자의 발아 특성 구명 ... 11
• 제2절 초분광 영상처리를 이용한 종자 발아 예측 기술 ... 15
• 제3절 라만 분광법을 이용한 종자 발아 예측 기술 ... 27
• 제4절 우량종자 선발 시스템 개발 ... 39
• 제 4 장 연구개발목표 달성도 및 대외기여도 ... 43
• 1절 목표대비 대외 달성도 ... 43
• 2절 정량적 성과 ... 43
• 제 5 장 연구개발결과의 활용계획 ... 45
• 제 6 장 연구개발과정에서 수집한 해외과학기술정보 ... 46
• 제 7 장 기타 중요 변동사항 ... 46
• 제 8 장 국가과학기술종합정보시스템에 등록한 연구장비현황 ... 46
• 제 9 장 참고문헌 ... 47