An intelligent System for Soil Classification using Supervised Learning Approach

Abstract

Agriculturist or farmers collect soil sample that are later analyzed for proper classification. This conventional procedure is labour intensive, time consuming and expensive. In this research work, an attempt was made to develop an intelligent system that can identify different types of soil in a particular location using the available hyperspectral data at such location with supervised learning approach. The system was developed using fuzzy –C means to identify the cluster centre. The cluster center was used as an input to train KSOM and generate soil prediction map as an output. ANFIS was eventually used to identify each class of the soil using the soil predictor map as an output during the training stage. The system was implemented using R programming Language.

Keywords: Hyperspectral data, C-Means Clustering, KSOM, ANFIS, Supervised Learning, Intelligent System,