对于大尺寸高精密光学元件，不仅要对光学元件表面低频面形精度和高频粗糙度进行控制，还需要严格限制中频误差，以保证其使用性能和稳定性。为了确定光学元件的不合格区域并指导其返修，引入经验模态分解(EMD)和Wigner分布(WVD)函数方法，通过理论分析确定该方法与功率谱密度函数间的关系，实现对光学元件表面中频误差的辨识与定位。实验结果表明：EMD-WVD方法不仅可以识别分布在实验光学元件表面15~27 mm空间频率为0.1 mm-1的中频误差，还可以减小多分量信号所引起的空间频率为1.0~1.5 mm-1的交叉项干扰，提高中频误差辨识的准确率。

The mid-spatial frequency error of high precision optical surface is crucial to the performance of high-energy laser system. To assure the performance and stability of the system, the low frequency error and high frequency error of the optical lens surface must be strictly controlled, the mid-spatial frequency error must also be strictly limited. In this paper, the relationship between empirical mode decomposition-Wigner-Ville distribution (EMD-WVD) method and power spectrum density is analyzed, and EMD-WVD diagnosis method is applied to identifying and locating the mid-spatial frequency error of the precision optical surface. The experimental results show that, with EMD-WVD method, the mid-spatial frequency error distribution of optical surface can be located in 15-27 mm and the spatial frequency is 0.1 mm-1, the interference of cross-term caused by multi-component signal whose spatial frequency is about 1.0-1.5 mm-1 can also be reduced, which enhances the identification accuracy of mid-spatial frequency error.