Binary class and multi-class plant disease detection using ensemble deep learning-based approach. (26th October 2022)

• Record Type: Journal Article
• Authors: Sunil, C.K.; Jaidhar, C.D.; Patil, Nagamma
• Is Part Of: International journal of sustainable agricultural management and informatics. Volume 8:Number 4(2022)
• Page Start: 385
• Page End: 407
• Publication Date: 2022-10-26
• Subjects: deep learning -- ensemble model -- machine learning -- plant pathology; Sustainable agriculture -- Management -- Periodicals; Sustainable agriculture -- Data processing -- Periodicals; Agricultural productivity -- Management -- Periodicals; Agricultural productivity -- Data processing -- Periodicals; Agricultural resources -- Data processing -- Periodicals; Agricultural systems -- Management -- Periodicals; Agricultural systems -- Data processing -- Periodicals; 338.1
• Languages: English
• ISSNs: 2054-5819

Abstract

Providing food for the exponentially growing global population is a highly challenging task. Owing to the demand and supply gap may diminish food production due to diseases in plants, such as bacterial disease, viral disease, and fungal diseases. Early recognition of such diseases and applying an appropriate pesticide or fertiliser can improve crop yield. Accordingly, early plant disease detection necessitates continuous crop monitoring from its initial stages. Recently some research works have been proposed as remedial measures. However, such methodologies utilise costly equipment that is infeasible for small-scale farmers. Thus, there is a need for a cost-effective plant-disease-detection approach. This study embellishes the challenges and opportunities in plant disease detection. Correspondingly, this research proposes an ensemble deep learning-based plant disease diagnosis approach using a combination of AlexNet, ResNet50, and VGG16 deep learning-based models. It effectively ascertains plant diseases by analysing the plant leaf images. A broad set of experiments were conducted using different plant leaf image datasets such as cherry, grape, maize, pepper, potato, strawberry, and cardamom to evaluate the robustness of the proposed approach. Experiential results demonstrated that the proposed approach attained a maximum detection accuracy of 100% for binary and 99.53% for multi-class datasets.