Study On Diagnostics Of Nutrient Elements Deficiencies In Establishment Planting Crop Using Hyper-spectral Imaging Technology

With the rapid development of modern agriculture, China has the world's largest area of facility agriculture, with a total of 3.3 million ha. As the establishment planting crop's short growth cycle, large yield and fertilizer requirements, leads to high-cost and fast-replenish in soil fertility. This used to cause nitrogen (N), phosphorus (P) and magnesium (K) deficiencies in the growth of cultivated crops, affect the yield and economic profits. Fast and accurate diagnostics of nutrient deficiency in cultivated crops has very important significance of increasing yield and the income of framers, the development of modern agriculture. In order to overcome the disadvantages of existing methods for diagnostics of nutrient deficiency in cultivated crops, hyper-spectral imaging technology was used to characterize physiological and morphological symptoms resulting from nutrient deficiency. We hope to obtain a new method for diagnostics of nutrient deficiency in cultivated crops. The main research works in this paper has been summed as following:(1) Cultivation of nutrient deficient samples. N, P, K, Mg deficient mini-cucumber plants and control plants were grown under non-soil conditions. Kjeldahl method, spectrophotometer method and atomic absorption spectrometry were used to determine N, P, K and Mg concentrations of mini-cucumber plants in N, P, K, Mg deficient and control groups. Results showed that:N, P, K and Mg concentrations for new, middle and old leaves were in a decreasing order. In comparison with N, P, K and Mg concentration of older leaves in control group, N, P, K and Mg concentration of older leaves in nutrient deficient groups decline markedly. This indicated plants in N, P, K and Mg deficient groups entered into nutrient deficient states. The successful cultivation of nutrient deficient samples laid a good foundation for further analysis.(2) Analysis of nutrient deficient symptoms and its distribution. Leaves at every single node in plant of N, P, K, Mg deficient and control groups were used as analytical samples. Chlorophyll, lutein and carotene concentrations of leaf sample were determined by high performance liquid chromatography (HPLC); image information was recorded by a color camera, then the nutrient deficient symptoms and its distribution in whole plant and leaf were analyzed. Results showed that:N deficiency caused to chlorophyll decreased in the whole older leaves, P deficiency did not change the leaf chlorophyll content, but resulting in the occurrence of small chlorotic spots, K deficiency caused to chlorophyll decreased in edge of older leaves, Mg deficiency caused to chlorophyll decreased between vein of older leaves. The analytical result figured out how to collect analytical sample, which technology to be used, and what kind of deficient symptoms to be extracted for further analysis.(3) Determination of chlorophyll concentration distribution map on mini-cucumber leaf. Spectra of 120 cucumber hyper-spectral images were extracted; interval partial least-squares (iPLS), synergy interval partial least-squares (siPLS), genetic algorithm interval partial least-squares (GA-iPLS) and genetic algorithm- simulated annealing algorithm-interval partial least-squares (GA-SA-iPLS) were used to select the efficient wavelength regions for chlorophyll concentrations (No.10,11,13,17,18), Principal Component Analysis (PCA) and Independent Component Analysis (ICA) were used to extract information from the selected wavelength regions, then the optimal chlorophyll calibration model was obtained (Rp=0.8769, RMSEP=2.42mg/g), spectral of every pixel in hyper-spectral images were computed and chlorophyll content of every pixel was obtained according to the calibration model. Finally, the chlorophyll distribution map was estimated. In comparison with chemical and spectral methods, the hyper-spectral technology could determine the chlorophyll concentration in the whole leaf.(4) Diagnostics of N, K and Mg deficiencies base on chlorophyll concentration distribution map. Chlorophyll concentration distribution map of N, K, Mg deficient and control leaves were determined. According to their chlorophyll concentration distribution maps, new diagnostics methods for N, K, and Mg deficiencies were proposed, the diagnostic rate for N, K, and Mg deficiencies were 97%,90% and 90%; respectively. The hyper-spectral diagnostic method overcomes the disadvantages of chemical and spectral diagnostic methods.(5) Diagnostics of P deficiency base on near infrared hyper-spectral image. Leave in P deficient and control group were used to collect near infrared hyper-spectral image; PCA and ICA were used to extract characteristic images. Results showed that the first ICA image could be used to extract P deficient symptoms; furthermore, this method could diagnose P deficient symptoms prior to the occurrence of responses to P deficiency that can be observed visually.(6) Development of nutrient deficient database soft for cultivated crops. This soft could record the kind of crop, varieties, kind of deficient nutrient, deficient image and spectra. It could be used for digital description of symptoms caused by nutrient deficiencies.In this study, new methods for non-destrctive diagnostics of nutrient elements deficiencies are proposed, and there is also of great significant in improve the level of facility agriculture in China.