Modified Artificial Hummingbird Algorithm with Deep Learning for Pest Detection through Leveraging Pattern Recognition-based Fusion Approach

The limitations of manual inspection and simple algorithms highlight the need for advanced insect detection systems in agriculture. The study presents the Modified Artificial Hummingbird Algorithm with Deep Learning (MAHADL), a significant breakthrough in pest detection. MAHADL combines the speed and efficacy of the Artificial Hummingbird Algorithm (AHA) with deep learning models such as Recurrent and Convolutional Neural Networks. Combining pattern recognition with the fusion approach improves pest detection accuracy while reducing false positives and negatives. Accurate pest identification is essential for effective agricultural management, which MAHADL provides through data collection, preprocessing, and algorithmic integration. MAHADL outperforms existing systems in terms of accuracy (0.94), recall (0.95), precision (0.92), computational economy (12 hours of training time, 0.5 seconds/image inference time), and decrease in false positives and negatives. The innovative technique provides enhanced agricultural output and sustainability, representing a significant leap in intelligent pest management.