This paper presents a comparative study on vehicle detection from UAV-captured aerial images using two prominent object detection algorithms: the single-stage detector SSD (Single Shot Multibox Detector) and the two-stage detector Faster R-CNN (Region-based Convolutional Neural Networks). The study aims to evaluate the performance of these algorithms across metrics such as precision, recall, mean Average Precision (mAP), F1 score, and detection time, providing insights into their suitability for UAV-based traffic monitoring applications. Results show that Faster R-CNN achieves higher detection accuracy, with precision, recall, and mAP values of 0.851, 0.813, and 0.826, respectively, outperforming SSD’s corresponding values of 0.770, 0.744, and 0.751. However, SSD’s detection speed (26.5 ms) is significantly faster than Faster R-CNN (63.2 ms), making it more suitable for real-time applications where immediate data processing is critical. Figures illustrate scenarios where Faster R-CNN exhibits superior performance, especially in complex scenes with dense vehicle presence, while SSD proves effective in simpler environments. This study highlights the trade-offs between accuracy and speed in object detection and suggests potential applications for each model in UAV-based traffic monitoring.