A Novel Method for Predicting Plant Leaf Disease Based on Machine Learning and Deep Learning Techniques

Abstract

Plants are valued for their practical purposes (food, medicine, etc.) and are thus considered indispensable. Diseases that attack the plant's leaves might strike at any time among plantings, wreaking havoc on the crop and its economic value. Consequently, the ability to recognize leaf diseases is extremely important in the agricultural sector. Nonetheless, it calls for a lot of work, additional time for planning, and an in-depth understanding of plant pathogens. Both “deep learning (DL) and Machine Learning (ML)” algorithms for diagnosing leaf diseases have been created and evaluated by several researchers, with generally favorable results for both categories of techniques. However, they still have certain issues when it comes to covering large regions to detect leaf diseases. We offer a new method for disease prediction in leaves, which we name Multi-Gradient Deep Convolutional Neural Network with Adaptive Support Vector Machine (M-D-C-A-S). We were able to cover he largest area with the greatest accuracy, and our technique outperformed other more conventional approaches. To evaluate the efficacy of the prediction of leaf diseases, a variety of measures are used, like accuracy, precision, f1-score, and recall. Experiments that predicted plant leaf diseases demonstrated the viability and efficacy of our technique.