| With the rapid increase of the world’s population, human demand for agricultural yield and quality have been greatly increased. And in order to increase crop yield, the use of fertilizers and pesticides had serious overload bearing capacity of the soil, causing many serious environmental problems, such as: soil harden and fertility decline. So the implementation of precision agriculture is particularly important, and precision fertilization is an important part of precision agriculture. Accurately obtaining the accuracy content of the soil necessary plant growth substances nutrients is the premise of precision fertilization, and then in the process of fertilization, on the basis of the chemical fertilizers accurate data and pesticides to apply fertilizer by scientific proportioning, implementation with minimal chemical inputs to obtain the highest production efficiency.At the present stage, for the measurement of soil nutrient content in basically has two methods: one most widely used the most accurate method is chemical analysis, the other is near infrared spectroscopy with newly developed in recent years but is not mature. Chemical laboratory examination method exists of detecting low efficiency, high cost, complex and higher requirement to operating personnel, and near infrared spectroscopy is currently only stay in the laboratory stage, but it has drawbacks of the near infrared spectrum instrument is expensive, lack of special testing instruments, lead to near infrared detection method cannot realize online realtime application field in agricultural production.In this study adopt the near infrared spectroscopy to realize the soil nutrient online accurate detection in the field, and in view of three key technology such as the near infrared detection of soil nutrient terminal design, the near infrared sensor design, spectrum data processing method was studied. The main content of the research are as follows:1. The soil nutrient on-line detection terminal system was studied based on near infrared technology. During the study, according to the theory of testing instruments designed each key structures and components, completed instrument light path structure, the design of hardware circuit and software program. In the design of light path structure, determined the detection methods of diffuse reflection principle, adopted the Czerny – Turner cross gambrel light path structure light, and through the peaceful face blazed grating dispersion to realize from the near-infrared compound light to near infrared light conversion. Realized light signal transform to electric signal by photoelectric conversion circuit, obtained the near infrared spectrum data of the soil.2. Based on the MEMS technology, designed a PbS/CuPc near-infrared light electricity sensor chip. Through chemical bath deposition method prepared lead sulfide type photoconductive sensor on glass substrate, by vacuum evaporation coating prepared CuPc photoelectric triode. Aiming at the process steps preparation of lead sulfide synthesis and chemical bath deposition of lead sulfide film, chip preparation process was studied, implemented the preparation of two different principle and photoelectric sensor of light-sensitive materials on a glass substrate. Broaded the scope of the NIR response spectra to replace the origin multiple sensors to simplified the instrument structure. And key parameters such as the sheet resistance, the I-V characteristics, temperature characteristic and microstructure characterization was tested.3. Complete the black soil samples, determined soil organic matter and total nitrogen content in the samples by the chemical detection method. Soil library of organic matter and total nitrogen content was established. Applied a variety of spectral data preprocessing methods to preprocess black soil samples near infrared spectroscopy spectra data, by comparing the effect of different methods of pretreatment, found a suitable pretreatment method.4. After the implementation of the spectra data pretreatment, through different methods of prediction model respectively to establish the prediction model of soil organic matter and total nitrogen content, and through two parameters: the correlation coefficient and root mean square error of the model to contrast the effect of different modeling methods, to find a suitable soil organic matter and total nitrogen content prediction model establishment method.The results show that using near infrared spectroscopy can realize soil organic matter and total nitrogen content prediction, and there is a bigger relevance between prediction value and accurate value laboratory using testing method, the highest correlation coefficient can reach 98.8% and 95.7% respectively, prediction effect is good, it can realize real-time measurement in the field of the soil organic matter and total nitrogen content. |
PDF Full Text Request