LDDNet: An Efficient Neural Network Model for Plant Leaf Diseases Detection.

The agricultural sector has a prominent role in contributing to the economy of many countries. Over the decades, production in the agricultural sector has decreased due to various factors such as leaf diseases, an overdose of chemical medication, natural disasters, and climatic changes. Majorly, the impact of plant diseases set a huge loss to the farmers compared to other kinds. Consulting an expert is a time taking and expensive process. Many machine learning and advanced deep learning algorithms are being implemented to identify diseases, more accurately, using the infected plant leaf image. The objective of this paper is to introduce a lightweight leaf disease detection Neural Network (LDDNet) that should be able to distinguish between diseased and healthy plants. The dataset contains 33 classes of different diseased and healthy plant leaves images, where each class has 1,680 training and 420 validating images. The accuracy obtained by the proposed LDDNet model is 99.30%. Since the performance of the model is high, it can be implemented in daily life to monitor plant diseases to have a healthy crop yielding. [ABSTRACT FROM AUTHOR]