Deep Learning Approaches for Plant Disease Detection: A Comparative Review
Mayank Agarwal, Ashish Kotecha, Atharva Deolalikar, Ritika Kalia, Ram Kumar Yadav, Achamma Thomas
Abstract
Crop production and growth are crucial factors that have an impact on the agricultural industry, as well as the farmer economically, socially, and in every other manner. Crop failure has several key causes, one of which is plant diseases. Plant diseases can be identified by a variety of signs, including infections, color changes, damaged leaves or stems, abnormal floret, stem, leaflet, bud, or root growth, among others. In addition, leaves exhibit disease indicators such as stains, dryness, premature dropping, etc. The research on using different Deep CNN architectures to detect illnesses of plants from photographs of leaves is reviewed in this paper. It also offers a comparison of models of CNN that were trained on leaf images to recognize and categorize the diseases of plants. Along with the benefits and drawbacks of the various approaches that were described in the prior research, it gives an overview of the performance measures and database used to assess the effectiveness of existing models.