Deep Learning-Based Web Application for Flower Recognition and Status Monitoring

This work presents a Flower classification and status monitoring web application based on learning for diverse flower species. This classification holds immense importance for farmers as it provides valuable insights into plant species, facilitating informed decision-making for crop management, pollination strategies, and pest control, ultimately enhancing agricultural productivity and sustainability. The work adopted the DenseNet169- based Convolutional Neural Network model. The results show an accuracy of 94%, recall of 92%, precision of 93%, and f1-Score of 93% using the Adam optimizer. A custom data augmentation pipeline, incorporating various image transformations, is introduced to enhance the model's robustness. A key contribution of this research is developing a user-friendly web application seamlessly integrating the trained model, allowing farmers to identify flower types within their fields and comprehend their agricultural significance.