Characteristic bands extraction method and prediction of soil nutrient contents based on an analytic hierarchy process

Abstract

The selection of characteristic bands of soil nutrients is crucial for the prediction of soil nutrient contents. In this paper, a measurement method based on the analytic hierarchy process (AHP) and reflectance spectrum is proposed to realize the extraction of characteristic bands and the prediction of soil nutrients. Firstly, a series of alternative bands are obtained by mathematical transformations and correlation analysis from the reflectance spectral curves of 30 soil samples. Then, to objectively reflect different dependencies of spectral bands on the soil nutrients, AHP is adopted for weight analysis, and the particle swarm optimization algorithm (PSO) algorithm is used to correct the expert scoring matrix. Via the weight allocation, weight normalization and weight threshold, the alternative bands meeting the sum threshold of weights > 0.5 are finally selected as the characteristic bands of soil nutrients, which are at 484 nm, 512 nm, and 760 nm for total nitrogen (TN), and at 384 nm, 736 nm, and 748 nm for available phosphorus (AP). Finally, three inversion models established by the characteristic bands are used to predict soil nutrient contents, which demonstrated the feasibility of the analytic hierarchy process for the extraction of characteristic bands and the prediction of soil nutrients.