컬러 영상처리에 의한 시설재배지 토양의 생물 물리적 환경변수 추정

• Kim, H.T. (Dept. of Bio-mechatronic Engineering, Sungkyunkwan University) ;
• Kim, J.D. (Dept. of Life Science, Hanyang University) ;
• Moon, J.H. (Dept. of Bio-mechatronic Engineering, Sungkyunkwan University) ;
• Lee, K.S. (Dept. of Bio-mechatronic Engineering, Sungkyunkwan University) ;
• Kang, K.H. (Dept. of Food and Life Science, Sungkyunkwan University) ;
• Kim, W. (Dept. of Bio-mechatronic Engineering, Sungkyunkwan University) ;
• Lee, D.W. (Dept. of Bio-mechatronic Engineering, Sungkyunkwan University)
• Published : 2003.08.01

Abstract

This study was conducted to find out the coefficient relationships between intensity values of image processing and biological/physical parameters of soil in greenhouses. Soil images were obtained by an image processing system consisting of a personal computer and a CCD earners. A software written in Visual C$\^$++/ systematically integrated the functions of image capture, image processing, and image analysis. Image processing data of the soil samples were analyzed by the method of regression analysis. The results are as follows. For detecting soil density of unbroken soil samples, the highest correlation coefficients of 0.82 and 0.84, respectively were obtained fur R-value and S-value among image processing data while it was 0.97 for G-value. Considering the relationship between biological characteristics and image processing data of soil in greenhouse, the correlation was found generally low. For pH of unbroken soil sample, the correlation coefficients were found 0.87, 0.85, and 0.94, respectively with G, I, and H values of image processing data. In the case of bacteria, any correlation was not found with the image processing data For Actinomyctes, they were 0.86 and 0.85, respectively with G-value and B-value of image processing data showing high correlation coefficient compared to the other variables. The correlation coefficient between Fungi and H-value was shown 0.88, the highest among the variables higher than 0.8 while the other variables showed low correlation. For broken soil samples from greenhouse, the relation between biological parameter and image processing data were rarely shown in this study. The results of this study indicated that most of correlation coefficient between the variables were usually lower than 0.01. Accordingly, it was assumed that the soil should be used without broken to fairly estimate biological characteristics using CCD camera.

References

1. ASAE Paper No. 964014 Optimizing the compostion process for miosture removal:Treroetical analysis and exterimental results Richad,T.S.;H.L.Choi.
2. Trans. of the ASAE. v.34 Evaluation of reflectance methods for soil organic matters Sudduth,K.A.;J.W.Hummel https://doi.org/10.13031/2013.31816
3. 농업과학기술원보고서 NIR을 이용한 부산물 퇴비의 비파괴 분석법 연구 남재작
4. 토양미생물실험법 토양미생물연구회편
5. 土讓の物理性 v.59 近亦外反射光分光法による土讓中の定量分析 谷山一郞;芝山道郞
6. 토양미생물연구회편. 1992. 토양미생물실험법. 동경월드사이언스 발행.
7. 北陸作物學會報 v.78 近亦外反射光分光法中の全窒素, 全炭素および抽出有機態窒素の測定 井上直人;晞井貴之