Image-Based Stress Detection in Maize Crop Using Deep Learning

Maize is a major cereal crop that plays a vital role in the agriculture driven economy of Pakistan. However, environmental factors and climate change introduce multiple stress conditions that significantly reduce crop yield. Early and accurate detection of crop stress is essential for improving agricultural productivity. This study proposes an image based deep learning framework for stress detection in maize plants. A dataset of maize plant images was collected from fields and augmented with publicly available data to ensure class balance. Six deep learning architectures were implemented and evaluated. Standard image preprocessing techniques were applied, and the dataset was divided into training, validation, and testing subsets. Experimental results demonstrate that the DenseNet model outperforms other architectures in terms of accuracy of approximately 93.5%. The proposed framework highlights the feasibility of a cost-effective and efficient system for maize stress detection, with potential applications in precision agriculture. Furthermore, to enhance practical usability, a web application was developed to enable real-time stress detection through user uploaded maize images. A web application was developed to enable stress detection through user uploaded maize images. The proposed framework highlights the feasibility of a cost-effective, scalable, and accessible solution for maize stress detection, with significant potential for applications in precision agriculture.

Poster