Research On Diagnosis Mechanism And Model Of Soil’s Water And Salt Content And Tomato Plants’ Moisture Based On Hyperspectral Imaging Technique

Ningxia Hui Autonomous Region is located in the upper reaches of the Yellow River in western China, and is a typical continental semi-humid semi-arid climate. The special economic crop industry in our area is one of the most promising agricultural growth projects in Ningxia, and the exact supply of soil water and fertilizer directly affects the high yield and quality of crops. Therefore, how to obtain cheap,quick and labor-saving ways to obtain the dynamic information of water and salt content in saline soil in arid and semi-arid areas is of great significance to the rationalization of saline soil.Based on Vis-NIR and NIR hyperspectral imaging, combined with stoichiometric methods, the dynamic monitoring of soil’s water and salt content and tomato plant’s water content were carried out.The rapid diagnosis of plant water deficit and detection mechanism of soil’s water and salt were studied in this study. The main findings include:(1)The variations of soil water content was different under different water conditions. For different salt content of soil, salt content in soil had great influence for redistribution of soil water. Irrigation of 0.2% salt concentration of soil column was better to control the migration of water in soil than irrigation of 2% salt concentration of soil column. The change rule of different moisture content, salt content of soil was analyzed, and set up four kinds of water content mathematical models between surface soil and deep soil.(2)The reflectivity of soil spectrum decreased with the increase of soil water content. When the amount of water, the reflectivity of the spectral curve increased with the increase of soil water content.The modeling effect of the characteristic wavelengths extracted by different methods, different modeling methods, different spectral ranges,and characteristic wavelength and full band modeling were discussed, and the MLR model based on the SPA method in 900～1700nm was optimized. The characteristic wavelengths were 987, 1386, 1466, 1568, 1636, 1645 nm. The correlation coefficient and root mean square error of predict set of soil’s water content were 0.984,0.631,respectively.(3)The situation of soil evaporation is also different under the influence of the increase of salinity of soil. During different days, the reflectivity of soil spectrum increased with the increase of soil salinity. The difference of soils’ reflectance were smaller for the high salt content of soil, which provides the theory basis for qualitative criterion of soil’s salt content of intelligent remote sensing in the future.(4)The modeling effect of the characteristic wavelengths extracted by different methods, different modeling methods, different spectral ranges, and characteristic wavelength and full band modeling were discussed, and the PLSR model based on the p coefficient method was extracted in 900-1700nm. The results showed that the characteristic wavelengths were 936, 996, 1016, 1136, 1151,1186, 1273, 1395,1425, 1458, 1535, 1642nm. The correlation coefficient and root mean square error of predict set of soil’s salt content were 0.949, 2.914g/kg, respectively.(5)The relationship between spectrum information and water content in tomato’s leaves and biological control mechanisms of salt-water coupling were analyzed. The modeling effect of the characteristic wavelengths extracted by different methods, different modeling methods, different spectral ranges, and the characteristic wavelength and the effect of full band modeling were discussed.The PLSR model based on SPA method was extracted in 900-1700nm, and the characteristic wavelength was 918, 981, 1029, 1387, 1652nm. The correlation coefficient and root mean square error of predict set of leaves’ water content were 0.9,0.614, respectively.(6)The models of soil’s water and salt content and tomato’s water content were built by hyperspectral imaging technique.The relationship among deep soil, surface soil, tomato’s canopy and hyperspectral were established, which lay the foundation for soil’s water and salt content in Ningxia and rapid remote sensing detection of plant leaves.