Open AccessA Mathematical Model to Estimate Soil Parameters using Wireless Sensor: An Efficient Way of Data Collection in Precision Agriculture
Author(s) -
Avishek Jana,
Arindam Roy
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.b6355.129219
Subject(s) - environmental science , water content , soil science , soil water , ground truth , precision agriculture , data collection , agricultural engineering , computer science , agriculture , statistics , mathematics , engineering , machine learning , geotechnical engineering , ecology , biology
Environmental and field parameters have high impact on agricultural productivity. The environmental parameters include temperature, humidity, rainfall and wind direction etc. whereas field parameters include salinity, nutrient content, oxygen levels, soil type, soil PH and soil moisture etc. Among them, Soil pH and moisture are highly correlated which can affect soil salinity, nutrient level and soil conductivity. So, these two parameters need to be measure precisely to take management decisions. Till now the process which is applied to measure soil parameters are entirely depends on laboratory testing of soil sample. This process has some overhead like availability of laboratory, manpower and cost. To overcome these challenges, we developed a virgin sensor based automated data collection technique which maintains the agricultural productivity with sustainable development. But practically sensor-based data retain some error which defers from laboratory result. In this paper, we have developed a new efficient technique to reduce the error in calculating pH value using sensors. In this research work, ten different types of soils are tested in laboratory to get the actual (pH level in soil. These values are considered as the ground truth for the experiment. The pH values of all the ten types of soil are collected from the field using wireless sensor. Our proposed mathematical model reduces the error to 0.01 between collected values and ground truth values with back propagation method based on soil moisture, environmental temperature and humidity. Here, the proposed model is empirically tested by taking some real field data values.