Design and Experiment of Portable Soil Organic Matter Detector Based on Spectroscopy Principle

In order to obtain the information of soil organic matter content quickly and non-destructively, a portable instrument for measuring soil organic matter content was designed based on spectroscopy principle. The detector was mainly composed of mechanical part, optical path system and control part, in which the mechanical part provided platform support for the detector, while the optical path part consisted of light source, sapphire glass, filter and photoelectric detector, and the control system realized the collection and processing of soil measurement signals. When the portable soil organic matter detector worked, the light emitted by the light source irradiated the surface of the soil to be detected, the diffuse reflection light was filtered by the optical filter, and then converted into an electrical signal by the photoelectric converter, and then the reflectivity at each sensitive wavelength was calculated by the signal processing unit, and the content of soil organic matter was detected by measuring the spectral reflectivity. The spectral data of soil and the measured values of soil organic matter content in Beijing Shangzhuang Experimental Station were collected. After preprocessing the spectral data, the processing results of four wavelength screening algorithms, CARS, MCUVE, MWPLS and Random Frog Leaping, were compared, and the partial least squares and random forest prediction models of soil organic matter content were established. The results showed that the random forest model based on four characteristic wavelengths selected by CARS algorithm had the best prediction accuracy, with the modeling set R2 being 0.923 and the prediction set R2 being 0.888. The CARS-RF model was embedded into the organic matter detector system. The experimental results showed that the correlation coefficient between the measured value and the standard value of the detector reached 0.891. The developed detector had high precision and can quickly detect the content of soil organic matter. © 2021, Chinese Society of Agricultural Machinery. All right reserved.