Review on Various Deep Learning Methods Adopted to Improve the Plant Leaf Disease Classification in Precision Agriculture

Precision agriculture achieves substantial development to increase the agricultural yield. Recent studies in perception helps to enhance the precision agriculture by the earlier finding of diseases and infection in plants. In this method, we conduct a review in various Imaging techniques for a crop monitoring method. Here various image processing applications are considered for plants classification, based on the harshness of disease or early detection of stress and classification accuracy. The main focus of this method is the utilization of hyperspectral imaging and classification of plant health and its ability to classify the disease present in the plant leaves. The study in general considers various deep learning strategies adopted to classify the hyper-spectral images. It further studies the classification accuracy of the deep learning methods like deep Convolutional Neural Network (CNNs), Deep Neural Networks (DNNs) and deep transfer learning. The study further provides the inferences related to the classification of hyper-spectral imaging using deep learning algorithms and its pitfalls like falling at local minima and other related challenges during classification.