Aarhus University Seal / Aarhus Universitets segl

Lashya P Marakkala Manage


Lashya P Marakkala Manage


Measuring soil properties by using Visible Near Infrared Spectroscopy

Project Description:

Visible and Near Infra-Red spectroscopy is recognized as fast and reliable technique to overcome some of the limitations in conventional laboratory methods to determine different soil properties. This method is accomplished to provide spectral information of the basic composition of soil, organic matter, minerals, water content, particle sizes, color and its structural qualities. The spectral data that combined with chemometric techniques provide better prediction performance to determine soil properties in lab scale as well as field scale.

However, the accuracy of spectroscopy prediction can be affected by external factors such as surface soil condition, soil temperature, aggregation of soil, particle sizes and soil moisture content. Soil moisture effect is predominant due to the fact that water has strong absorption bonds in the vis-NIR region and those bands may interfere with absorbance of other soil properties and degrade the prediction performance.

Therefore, the aim of the project is to study the moisture effect on soil spectra and vis-NIR spectroscopy prediction for soil properties basically soil organic carbon and soil texture. At the end, the study will focus to find appropriate spectral algorithm to remove moisture effect from spectra and to improve the accuracy in prediction. On the other hand, the awareness of minute variability in soil particle density may important for properties that are estimated by calculation using particle density. Indirectly, it may help to understand the amount of water or air content in soil and their movements in soil. Therefore, finding a precise, fast and consistent method to estimate soil particle density would be helpful in all soil applications. Therefore, this study will find out the feasibility of using Vis/NIR spectroscopy to infer soil particle density.

Vis alle (4) »

ID: 79006869