Theories and Methods for Soil Nutrient Sensing

Abstract

Soil provides the essential nutrient elements for plant growth. Soil nutrients include soil macronutrients such as nitrogen, phosphorus, and potassium, as well as soil micronutrients such as calcium, magnesium, sulfur, iron, and boron. Precision agriculture is a technology of applying precise and right amounts of inputs such as water, fertilizer, and pesticides at the right time to the crop for increasing the productivity and maximizing the yields. Therefore, it is necessary to obtain soil nutrient information quickly and accurately. Near-infrared spectroscopy (NIRS) with high-efficiency and nondestructive characteristics has great potential in soil nutrition detection. According to the NIR absorption of the hydrogen bonds, soil total nitrogen content and soil organic matter content can be estimated. Multiple linear regression, partial least square regression (PLSR), and principal component analysis (PCA) are commonly used to establish the estimation models of soil nutrient contents based on NIRS. Moreover, the modern algorithms of wavelet algorithm (WA), genetic algorithm (GA), uninformative variable elimination (UVE), support vector machine (SVM), etc., are used to reduce the multicollinearity of the NIR spectra to improve estimation accuracy. Laser-induced breakdown spectroscopy (LIBS) is a promising spectral detection technology with high sensitivity, fast speed, and the ability to measure multiple elements simultaneously. It can also be used to detect both soil macronutrients and micronutrients. At present, scientists have developed various forms of soil testing instruments based on spectral technology, such as portable, vehicle-mounted, and remote sensing devices. Through these devices, it is convenient to implement comprehensive, full-range, all-weather, and real-time soil sensing for soil and crop precision management.