Identification of Nutrient Deficiency Based on Leaf Image Data Using Machine Learning

The world population has been growing significantly during the last few decades. The demand for rice is constantly increasing with the ever-increasing global population, as rice crops are one of the primary food sources for humans. Increased fertilizer use in recent years has helped farmers to raise their rice production rapidly. Adding too much of fertilizers will have a negative effect on the soil. To boost the rice production, it is necessary to ensure that each paddy plant is healthy. This study makes use of a very practical method to identify nutrient deficiencies (NPK) in rice plants, using leaf imaging and machine learning techniques. The study proposes a framework in which the color and texture features of the image are extracted and then combined to get a single robust feature representation of the image. Classification of the image is done for identifying macro nutrient deficiencies using the Random Forest model, which when compared with Naive Bayes gave better accuracy.