Self-augmentation has been received increasing research interest recently to improve named entity recognition (NER) performance in low-resource scenarios. Token substitution and mixup are two feasible heterogeneous self-augmentation techniques for NER that can achieve effective performance with certain specialized efforts. Noticeably, self-augmentation may introduce potentially noisy augmented data. Prior research has mainly resorted to heuristic rule based constraints to reduce the noise for specific self-augmentation individually. In this paper, we revisit the two self-augmentation methods for NER, and propose a unified meta-reweighting strategy for these heterogeneous methods to achieve a natural integration. Our method is easily extensible, imposing little effort on a specific self-augmentation method. Experiments on different Chinese and English NER benchmarks demonstrate that our token substitution and mixup method, as well as their integration, can obtain effective performance improvement. Based on the meta-reweighting mechanism, we can enhance the advantages of the self-augmentation techniques without extra efforts.