The high proportion of new energy integrating into the power grid reduces its frequency regulation capability, leading to prominent frequency security issues. Considering frequency security constraints in the unit commitment becomes an important guarantee for system stability. In addition, the uncertainty of new energy generation also increases the system reserve scheduling pressure. To address this issue, a day-ahead unit commitment and reserve optimization model considering frequency security constraints and uncertainty of wind power is proposed. Firstly, the frequency security constraints are derived based on a frequency nadir prediction model, and then incorporated into the unit commitment model to guarantee the frequency security after power disturbances. Secondly, a distributed robust optimization framework is utilized to handle uncertainties on both the generation and load sides, coordinating thermal power and wind power to provide reserve capacity for the system. To reduce the conservatism of the model, the average Hausdorff distance is employed to construct an uncertain set that restricts the probability distribution of scenarios. Finally, the two-stage optimization model is decoupled into a master-problem and a sub-problem, which can be solved by column and constraint generation (C&CG) algorithm. The case study demonstrates that the proposed method is beneficial for balancing the economy and robustness of system operation, while also improving the frequency stability.