Drones equipped with various imaging sensors have become essential tools for urban monitoring, with applications spanning environmental change detection and traffic analysis. However, challenges such as small object sizes, viewpoint variability, and low resolution in high-altitude drone imagery limit the accuracy of object detection. In this study, we investigated the use of deep-learning-based super-resolution techniques to enhance object detection in drone imagery. The Super-Resolution Generative Adversarial Network (SRGAN) model was used to generated super-resolved imats at 2× and 4× scales to improve image quality. Objects were detected using the PaddlePaddle-You Only Look Once Enhanced-Small Object Detection (PP-YOLOE-SOD) algorithm, which enabled a comparative analysis of the object detection performance between original and super-resolved imagery. The findings indicate that 2× super-resolution significantly enhances the detection of small objects, such as pedestrians and two-wheeled vehicles, leading to improved recall and F2-scores. In contrast, 4× super-resolution reduced the detection accuracy. In this study, we demonstrated that super-resolution techniques can effectively address challenges associated with drone imagery at high altitudes, enhancing detection performance for small objects. However, the results underscore the importance of selecting appropriate resolution scales to avoid diminishing returns. These findings offer valuable insights for optimizing drone-based monitoring systems in urban environments, with implications for traffic management and object tracking under challenging conditions.