INTERNATIONAL JOURNAL FOR INNOVATIVE RESEARCH IN MULTIDISCIPLINARY FIELD

Abstract: In India, agriculture is an important industry and source of revenue. A variety of crops are cultivated for both financial gain and food. Among the most important crop grown in many parts of India is cotton. The state of Maharashtra is the top producer of cotton. However, many diseases usually affect cotton crops, especially during growth. As a result, there are considerable yield losses. Every living organism in this world requires a sufficient amount of a variety of proteins and nutrients for its average growth. Deficiency of any of the nutrients in the crop impacts average growth and, in turn, results in poor quality of yield and financial loss. Nitrogen is a crucial nutrient for the average growth of plants, and its deficiency can lead to various symptoms in cotton leaves, such as yellowishness on the leaf and stunted growth. Early detection of nitrogen deficiency is must in guaranteeing the best yield output and prompt management. In this research, the combination of image processing and DenseNet 121 is used to identify and diagnose the deficiency. This combination gives percentage estimates of nitrogen deficiency along with the classification of unhealthy leaves. The model delivers an F₁ score of 98% and validation accuracy of 98.26%.