Development of machine learning algorithms for internet of things based smart agriculture systems

Abstract

Food scarcity in developing countries can be reduced by increasing the yield in agriculture. Three major reasons for the low yield are the farmers are unaware of the fertility of the land, excess use of fertilizer as it is randomly applied in the field without soil testing and the use of old methodology in farming that has no monitoring system in the agricultural field. newlineFirst and foremost, the decision on crops to be planted in the agricultural land should be based on soil fertility. The erroneous decision of the crop for a particular land causes loss in agricultural production. Secondly, during the growth of the plants, it absorbs nutrients from the soil that leads to a lack of nutrients for further growth. Thirdly, the absence of checks regarding soil nutrients through traditional methodology adopted by farmers might cause improper usage of fertilizer. Fourthly, nutrients are lost due to environmental hazards like soil erosion. As the productivity of agriculture depends on the soil nutrients of the land, the research ponders on monitoring the soil nutrients in an automated, accurate and economical method newlineSoil nutrients are classified as macronutrients and micronutrients. Macronutrients are substances present in high proportion in the soil namely Nitrogen (N), Phosphorous (P), Calcium (Ca), Magnesium (Mg), Sulfur (S) and Potassium (K). Micronutrients that present in the soil are less in proportion and less significant for plant growth. Chlorine (Cl), Boron (B), Manganese (Mn), Zinc (Zn), etc., are microelements also known as trace elements. Major consumption of nutrients required for plant growth are N, P and K. Hence, the determination of soil nutrients level is significant for the farmer in all processes starting from crop plantation till harvesting. newline