The development of Intelligent Transportation Systems (ITS) has been accompanied by rapid advancements in sensors, wireless communications, cloud computing, and data science, making a significant contribution to reducing traffic congestion, optimizing traffic flow, and promoting the construction of smart cities. However, in practical applications, due to reasons such as sensor malfunctions, software issues, or network connection interruptions, the collected traffic data often has missing values, which negatively impact subsequent traffic management and decision-making processes. Therefore, effective recovery of missing traffic data becomes critical. This paper reviews some methods for missing traffic data imputation that have emerged in recent years and provides a systematic classification and in-depth analysis of these approaches. The methods are primarily categorized into two types: structure-based methods and learning-based methods. Each approach has its unique advantages and limitations and is suitable for different patterns of missing data. Moreover, the paper discusses common patterns of missing data, public datasets, and performance evaluation metrics, providing researchers with references for choosing appropriate experimental resources. Finally, this paper points out the challenges currently faced and suggests future research directions to promote the development of traffic data imputation techniques.