空中交通系统是复杂非线性系统，具有分形特征。为了更精确把握飞行轨迹在时空上的分布结构，探索分形方法在飞行轨迹挖掘领域的应用前景，首先采集了广州机场2023年10月11日的ADS-B数据，以向北运行的522条轨迹为研究对象；接着采用改进量规法计算了轨迹的分形维数；然后采用去趋势互相关分析方法计算了轨迹的多重分形谱，进而分析了轨迹的多重分形特征；最后通过实例分析，研究了分形特征量在异常轨迹识别问题中应用的可行性。算例结果显示，本文算例中，有99.43%的飞行轨迹具有分形特征，93.68%的飞行轨迹具有多重分形特征，识别出3条盘旋等待的飞行轨迹。研究结果进一步验证了空中交通系统的分形特征，也说明分形方法在异常轨迹识别方面具有一定的应用潜力，但该方法的准确性和普适性仍需深入研究。

The air traffic system is a complex nonlinear system with fractal characteristics. In order to more accurately grasp the spatial and temporal distribution structure of flight trajectories, and explore the application prospects of fractal methods in the field of flight trajectory mining, ADS-B data from Guangzhou Airport on October 11, 2023 was first collected, with 522 trajectories running northward as the research object. Subsequently, the improved division method was used to calculate the fractal dimension of the trajectory. Then, the multifractal spectrum of the trajectory was calculated using the detrended cross-correlation analysis method, and the multifractal characteristics of the trajectory were analyzed. Finally, through case analysis, the feasibility of applying fractal feature quantities in abnormal trajectory recognition problems was studied. The calculation results show that 99.43% of the flight trajectories in this article have fractal features, 93.68% of the flight trajectories have multifractal features, and three hovering waiting flight trajectories were identified. The research results further validate the fractal characteristics of air traffic systems and indicate that fractal methods have certain potential applications in identifying abnormal trajectories. However, the accuracy and universality of this method still need further research.