Neural Network-based Soil Detection and Classification
A. Mary Sowjanya, K. Sonali Swaroop, Sandeep Kumar, Arpit Jain
Abstract
Soil classification is the disintegration of soil sets to specific gatherings having like attributes and comparable behaviors. Practically many nations do product trading, in which those nations sending out higher horticulture products are especially rely upon the soil qualities. In this manner, soil quality recognition and classification are a lot of significant. Recognition of the soil kind assists with keeping away from horticultural product amount misfortune. This paper introduces a fully connected network (FCN), deep learning model-based recognition of the soil kinds. Soil classification incorporates steps like image acquisition, feature extraction, and classification. The proposed method produces an average accuracy of 97.2% with an average mean of 65.27 and average energy of 0.0298. The proposed model classifies peat, sandy Clay, Silty Sand, and Human clay soil types effectively. Keywords: Classification; Fully Connected Network; Deep Learning, Soil Detection, Soil Classification.