Radish Classification by using Hybrid Deep Learning Approach

In response to the increased need for precise and effective agricultural techniques, this study presents an innovative and integrated method of sorting radish varieties. The methodology employs a state-of-the-art fused model, integrating Convolutional Neural Networks (CNN) and Support Vector Machines (SVM), to discern between three key radish varieties: Pusa Chetki, Pusa Desi and Pusa Himani. The study is based on detailed analysis of a wide and variable sample dataset, which encompasses 12000 high-resolution radish images. This dataset does not only describe all the specificity about every radish variety but also guarantees good training and testing of classification models. Notably, CNN-SVM fused model has an amazing overall accuracy of 96.6% in the classification task. This approach not only sets a new standard of radish variety identification but also demonstrates how different computational approaches can be blended together to address challenging agricultural issues. The results highlight the great potential that the hybrid models have in optimizing the classification accuracy for certain crop varieties, providing a promising direction toward informed decision-making. This research enables future improvements in crop classification techniques for adopting precision agriculture and sustainable food production practices.