AUTHOR=Zhang Xiaodong , Duan Chaohui , Wang Yafei , Gao Hongyan , Hu Lian , Wang Xinzhong 

TITLE=Research on a nondestructive model for the detection of the nitrogen content of tomato

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.1093671

DOI=10.3389/fpls.2022.1093671

ISSN=1664-462X

ABSTRACT=China is a big country of tomato planting and consumption, and the protected tomato planting area accounts for 57% of the total area. At present, most facilities still follow the extensive management mode of attaching importance to water and fertilizer. The reason lies in the lack of advanced and applicable methods to quickly perceive crop nutrition information and the lack of scientific basis for water, fertilizer and environmental regulation. Terahertz technology can be used to obtain the internal nutrition status of crops because of its penetrability and fingerprint. In this paper, terahertz spectrum information of different nitrogen levels is obtained by using terahertz spectrum, and noise reduction of terahertz spectrum data is carried out by using S-G smoothing algorithm. The sample set is analyzed by using KS and RS respectively, and the KS algorithm is optimized to divide the sample set. SCARS, UVE and iPLS algorithms are used to screen the pre-processed terahertz spectral data. Based on the selected characteristic frequency band, the detection model of tomato nitrogen content based on terahertz spectrum was established by RBFNN and BPNN algorithm. The results showed that the accuracy of the model established by RBFNN algorithm is slightly higher. The model's RMSEC = 0.1322%, RMSEP = 0.1855%, Rc2 = 0.8714, Rp2 = 0.8463, and the quantitative analysis of tomato nitrogen was preliminarily realized.