Application of Real-Time Deep Learning in integrated Surveillance of Maize and Tomato Pests and Bacterial Diseases
Keywords:ML, CNN, Tomato Disease, Maize Pest, Smartphone App, Kenya, Farmer Extension, Precision Agriculture, Resource-Constrained Regions
With an emphasis on maize and tomato crops specifically, this research explores the creative fusion of computer vision (CV) and machine learning (ML) to address the enduring problem of pests and crop diseases impacting Kenya's crucial agricultural industry. This study aims to provide farmers with a reliable and accurate tool for identifying pests and diagnosing diseases by using a MobileNetV2-based model. An extensive dataset including photos of both healthy and sick crops was gathered, and preprocessing approaches, such as data augmentation, were used to improve the model's training procedure. The resulting mobile application for real-time digital imaging signifies a major advancement in agriculture technology as it provides farmers with rapid and accessible diagnostic capabilities in addition to professional advice. The transformational potential of CV and ML in transforming agricultural practices is highlighted by this research, despite faced constraints such as data limits and technical accessibility. This research proposes a future where these cutting-edge technologies serve as pillars fortifying food security, empowering farmers with actionable insights, and significantly mitigating crop losses in Kenyan agriculture. It does this by arguing for improved data collection strategies, addressing algorithmic biases, and encouraging technology adoption. This paper contributes an important step toward a more technologically advanced farming environment, where innovative solutions utilizing CV and ML methodologies promise to improve farmer livelihoods and protect Kenya's agriculture from widespread threats in addition to preserving crop health.