Object detection is of great importance in Computer Vision (CV) domain. As Artificial Intelligence (AI) technique developed rapidly, a lot of object detection methods and their improved version has been proposed. However, in order to be utilized efficiently, the object detection method is supposed to adapt to small object, complex environment, process in real-time, require less computing power and to be generalization. In this paper, we review the object detection of traditional methods, two-stage methods, one-stage methods and vision language models (VLM). Targeting airborne application, such as unmanned aerial vehicle (UAV), we proposed an improvement object detection method based on YOLOv11, with modification of the structure. Experiments are conducted to validate the modification and compare the result to the baseline.