Volume No. 12 Issue No.: 4 Page No.: 468-463 April-June 2018

 

PREDICTING SOIL PROPERTIES AND MOISTURE CONDITION USING REGRESSION ALGORITHMS IN HYPERSPECTRAL DATA ANALYSIS

 

Jain Mukul*, Yadav Ankit and Mohapatra S. N.

Centre of Remote Sensing and GIS, S.O.S. in Earth Science, Jiwaji University, Gwalior, Madhya Pradesh (INDIA)

 

Received on : September 30, 2017

 

ABSTRACT

 

In this paper an attempt has been made to predict soil properties and moisture condition by chemometric method using hyperspectral image. Chemometric method is very fast and efficient techniques to determine soil properties as well as moisture condition which uses the regression coefficient denoted by b .In this paper various regression algorithms are discussed and compared. Partial Least Square Regression (PLSR) techniques are used to determine different components of soil i.e. sand silt and clay and moisture content. We can finally compared data predicted using PLSR algorithm. The limitations and advantages and of these algorithms are compared and discussed. The regression model may become more stable if the number of soil samples are increased. The result of the study indicates that Hyperspectral data analysis has improved the ability in predicting soil properties.

 

Keywords : Chemometric analysis, Hyperspectral imaging, Partial least square method

 

 

BACK