Prediction and detection of nutrition deficiency using machine learning

Abstract

The plant producers have a hard time identifying nutritional inadequacies in their crops. The capacity to recognize these comprehensive nutritional deficiencies could help regulate crops properly. Using image processing, Convolutional Neural Network (CNN), the researchers were able to categorize and identify complete nutritional deficiencies in various cultivars. The prototypes would provide prescribed plant fertilizers once nutrient insufficiency was recognized. Iron (Fe), magnesium (Mg), potassium (K), nitrogen (N), calcium (Ca) and complete nutrition were examined. For classifying the image processing techniques were used to turn the images into grayscale & binary data. Using identification and prediction, CNN predicts complete nutritional deficiencies in the plant. CNN high accuracy of detection and diagnosis of nutrient deficits in different cultivars, according to the results compared with Artificial Neural Network (ANN) and DenseNet-121. The design has been tested and the results demonstrate a better way to classify and diagnose complete nutritional deficits in different cultivars.