ARCHIVES
Advancements in Deep Learning for Accurate Detection of Leaf Diseases in Agriculture
1 2 3 4 Students, Department of Electronics & Communication Engineering, Indira College of Engineering & Management, Pune, Maharashtra, India. 5 Assistant Professor, Department of Electronics & Communication Engineering, Indira College of Engineering & Management, Pune, Maharashtra, India.
Published Online: May-August 2024
Pages: 01-03
The "Leaf Sickness Location" framework tends to the basic test of plant illnesses in horticulture through the execution of a computerized arrangement utilizing profound learning procedures. In this far-reaching try, convolutional brain organizations (CNNs), explicitly DenseNet-121, ResNet-50, VGG-16, and Origin V4, are calibrated for productive and exact ID of plant illnesses. The task uses the Plant Village dataset, including 54,305 pictures across 38 plant infection classes, to direct a similar investigation of model execution. DenseNet-121 arose as the top-performing model, accomplishing an excellent 99.81% grouping exactness, outperforming other best in class models. The framework's approach decisively utilizes move figuring out how to conquer computational difficulties related with preparing profound CNN layers. This methodology, combined with the multi-class order system, demonstrates vigorous in taking care of different plant species and sicknesses inside each class. The outcomes feature the unrivaled effectiveness of move advancing in contrast with building models without any preparation, exhibiting the potential for genuine applications in farming. The framework's prosperity is credited to the cautious improvement of hyper parameters and the reception of cutting-edge profound learning strategies, offering a promising road for robotized and precise plant sickness identification, with suggestions for working on rural works on, limiting financial misfortunes, and guaranteeing worldwide food security.
