Crop Disease Detection and Intelligent Fertilizer Predictor (KEY IJP************762)

• Akshita Panditrao Dudhanikar,Sharmili Anand Burkule,Shruti Ravishankar Chakre,Prajakta Nagnath Pise,Vaishnavi Sudarshan Yalshette,Prof. Yoginath Kalshetty

Abstract

In agriculture, the impact of plant diseases and pests on crop yield and quality is significant. Identifying and managing these issues is crucial for sustainable agriculture. Digital image processing techniques have emerged as powerful tools in this regard, with recent advancements in deep learning surpassing traditional methods. This review delves into the application of deep learning technology for the identification of plant diseases and pests, a topic of growing interest among researchers. The review begins by defining the problem of plant diseases and pests detection and compares it with conventional detection methods. It highlights the substantial progress made in recent years, focusing on the superiority of deep learning over traditional techniques. The study categorizes recent research into three aspects based on network structure: classification network, detection network, and segmentation network. Each approach's strengths and limitations are meticulously summarized. Furthermore, the review introduces commonly used datasets and conducts a comparative analysis of the performance of existing studies. It provides valuable insights into the effectiveness of various deep learning methods in real-world scenarios. Building upon this analysis, the study identifies potential challenges in the practical application of plant diseases and pests detection based on deep learning. These challenges include issues related to dataset quality, model generalization, and real-time implementation. In addressing these challenges, the review proposes innovative solutions and research directions. These solutions encompass techniques for dataset augmentation, transfer learning, and domain adaptation to enhance model robustness. Additionally, the study suggests exploring interdisciplinary collaborations between agriculture experts and machine learning researchers to develop more holistic and domain-specific solutions. Finally, the review offers a comprehensive analysis of the future trends in plant diseases and pests detection based on deep learning. It anticipates the integration of advanced technologies such as edge computing and IoT devices to enable real-time monitoring and decision-making. The review underscores the need for continuous research and development to bridge the gap between theoretical advancements and practical agricultural applications. Keywords: Deep Learning, Convolutional Neural Network, Plant Diseases and Pests, Classification, Object Detection, Segmentation.