Optimizing energy efficiency in agriculture through machine learning-enabled plant disease management

The study looks into the crucial subject of optimizing agricultural energy efficiency through disease management provided by machine learning (ML). The main goal is to investigate how ML technology transforms disease detection, and lowering the need for energy-intensive agricultural inputs. Through the utilization of Gray level co-occurrence matrix (GLCM) technique, the research conducts a thorough analysis of 5 publicly available datasets of plant datasets initially to define the best dataset, while setting a criteria scale, to be used in the deep learning model training. Dataset 2 scored the best performance hence used as the input dataset for the deep learning. Four well-known deep transfer learning models—VGG16, GoogleNet, ResNet50, and DarkNet53—were put into practice and assessed due to their good degree of accuracy in classification tasks. VGG16 (96.7% accuracy) and Darknet53 (99.7% accuracy) appeared to be the best performing models in the test results. A subsequent theoretical estimation of energy efficiency viable from machine and deep learning modelling were further computed while utilizing the best performing model. In addition to enhanced crop yields, a 40% decrease is estimated to be realized, using the approach in this study, to establish a substantial energy efficiency in plant disease management. These results presented novelty in methods and results, while encouraging opportunities for precisely managing diseases to reduce energy use in agriculture. Further recommendations were provided for future work based on the processes realized. In conclusion, the study emphasizes how ML technologies have the ability to completely change sustainable agricultural methods and lessen dependency on inputs that need a lot of energy.