Optimization of YOLO11 for Military Aircraft Target Detection in Remote Sensing Images

Abstrak

Military aircraft target detection in remote sensing images is a critical requirement in the intelligence field for protecting the sovereignty of a region. However, the process of detecting aircraft using remote sensing faces many challenges, especially when dealing with small-sized military aircraft, complex backgrounds, and varying weather conditions. Accuracy and speed in aircraft detection are essential for making important decisions. Traditional methods are no longer sufficient to meet the demands of effective and accurate target detection. This study employs a deep learning approach that achieves high accuracy and speed in detecting military aircraft in remote sensing imagery. The method uses YOLO11, optimized by tuning several parameters such as learning rate, optimizer, and epochs. Parameter tuning with a learning rate of 0.001, RAdam optimizer, and 300 epochs resulted in the highest mean Average Precision (mAP) of 98.7% with a speed of 172 FPS. Additionally, the model achieved its highest speed of 303 FPS with an mAP of 94.1% using a learning rate of 0.01, NAdam optimizer, and 15 epochs. With proper tuning of parameter combinations, the accuracy and speed of object detection in remote sensing imagery can be optimized.

Identifier

• DOI: https://doi.org/10.31328/jointecs.v10i2.7401
• OpenAlex: https://openalex.org/W7125177598