The goal of the project is to create an interpretable ML/AI/DL system that can rapidly and precisely recognize pests and diseases that affect crops. This system examines cropped photos and outputs precise and transparent detection results by utilizing machine learning algorithms and computer vision techniques. The interpretability feature makes sure that the AI/ML/DL system delivers concise justifications for its predictions, enabling farmers and agronomists to comprehend the concerns found and decide on the best course of action for managing pests and diseases. Enhancing crop health management techniques and maximizing agricultural productivity are the goals of the research.
Pests and diseases have an adverse effect on crop yield and food security, which presents difficulties for farmers all over the world. Effective pest management and disease control depend on the prompt and precise detection of these threats. Traditional detection techniques frequently require specialized knowledge and are subjective and time-consuming. This research tackles these constraints by offering clear insights into the detection process through the development of an interpretable AI system. By better understanding how the system makes decisions, farmers and agronomists can safeguard their crops, reduce losses, and increase agricultural sustainability.
Several directions can be explored in future research on explainable AI systems for crop pest and disease detection. Future research may concentrate on improving the interpretability of the AI system through the incorporation of domain knowledge and expert guidelines. The accuracy and dependability of the system could also be increased by looking at the incorporation of multi-modal data sources, such as satellite imaging, weather data, and historical records. Exploring the interpretable AI system's scalability and deployment in other crop types and geographical regions can also result in a larger adoption and useful implementations in the agriculture sector.
Some of the skills of this research topic include:
