Soil moisture level prediction using optical technique and artificial neural network

This research describes the use of an optical system combined with artificial neural network (ANN) for wireless and nondestructive prediction soil moisture level. The former comprising near infrared (NIR) emitters wavelengths 1200 nm 1450 nm, a photodetector real time measurement in loams peats holding different amount water. There were 63 90 sets data from peats, respectively, used development dual stage-multiclass ANN model, wherein light attenuation (from system) was correlated percent destructive gold standard approach) pre-measurement stage. result revealed relatively good performance training NN regression, R, 0.8817 0.8881, satisfactory error 0.7898 1.172, respectively. testing on 50 new samples loam peat showed considerably high mean accuracy 92 % while 82 observed peats. study attributes poorer to detection resolution moisture, structure properties corresponding soil. work concluded that developed technology may be feasible future design improvement agricultural management.