Object Detection Algorithms Identifying Ceratitis Capitata Fruit Fly

This study aims to obtain an artificial intelligence model for monitoring the fruit fly Ceratitis Capitata. Last artificial intelligence technologies are proposed to solve this task. So, this research utilizes deep learning object detection using YOLO algorithm to identify, detect, and count male flies on Jackson Trap images. The above model integrates model based and data-driven tuning methods such as automated hyperparameter tuning, data augmentation, image processing, and synthetic data generation to improve the score. This study is significant due to the global threat of this fly to crops and expensive current monitoring methods. Institutions deploy networks of Jackson Traps to monitor flies, but this manual process is labor-intensive. The challenge of expensive technique was solved by implementing the model studied, and promising results were achieved. So, this model exhibited a mean average precision of 0.86 for training and 0.84 in validation.