An IoT-based soil analysis system using optical sensors and multivariate regression

Abstract

Food is the primary requirement for the survival of any living being on this planet. The rapid increment in the population is a major concern for adequate food production due to the depletion of agricultural land, which has turned into housing societies. However, agriculture is India's main business and primary income source for the farmers. The agricultural crop yield mainly depends upon the physical parameters of the soil, such as micronutrients and pH values. The main constraint in monitoring these parameters is the location of land at the far remote places and it takes enough time to test these parameters following the lab test process. The real-time analysis of all the parameters remained a big challenge for the farm owner, so the soil fertility level could not be sustained at the optimum level during most of the crop production cycle. This ultimately results in the average level of crop production and becomes a matter of chance since the soil fertility and other parameters barely suit the crop type under cultivation. This paper mainly focuses on developing an Internet of Things (IoT) based digital method to measure the availability of soil macronutrients and their pH using a color optical sensor TCS3200 and transmit those parameters to a long distance in case of unavailability of any telecommunication network. The paper also describes the deployment of Long Range (LoRa) units interfaced with ESP8266 for long-distance communication and uploading the entire information over the cloud platform, which will be displayed over the mobile using an API. The average accuracy of the proposed method in determining the soil macronutrients was 0.969 for phosphorus, 0.953 for nitrogen, 0.961 for potassium, and 0.921 for Soil pH.