The existing reconstruction methods for dam monitoring data have the problems of being unable to reconstruct in the non-complete dataset and the reconstruction accuracy is not high enough. Therefore, this paper proposes the dam monitoring data reconstruction method (DSVD) to realize the accurate reconstruction of dam monitoring data in non-complete datasets. The method first adopts the sorting method, which is different from singular spectrum analysis, to construct the dam monitoring matrix. Then, the singular value decomposition is performed, and the hard singular value thresholding algorithm is used to select the singular values to form the dam monitoring approximation matrix. On this basis, an accurate reconstruction model is constructed, and the reconstructed values are used to replace the missing values to realize the accurate reconstruction of dam monitoring data. Finally, the reconstruction method of dam monitoring data is applied to monitoring system of Longgang Reservoir in Chongqing, and the results show that the method can realize the reconstruction of dam monitoring data in non-complete datasets, and the accuracy is higher than that of other models.

中文翻译：

---

一种基于改进奇异值分解的大坝监测数据重构方法


现有的大坝监测数据重建方法存在数据集不完整无法重建、重建精度不够高的问题。因此，本文提出了大坝监测数据重建方法（DSVD），以实现对大坝监测数据在非完备数据集中的精确重建。该方法首先采用不同于奇异谱分析的排序方法构建大坝监测矩阵。然后，进行奇异值分解，采用硬奇异值阈值算法选择奇异值，形成大坝监测近似矩阵。在此基础上，构建了精确的重建模型，并利用重建的值替换缺失的值，以实现大坝监测数据的精确重建。最后，将大坝监测数据重建方法应用于重庆市龙岗水库监测系统，结果表明，该方法能够实现对不完全数据集中大坝监测数据的重建，且精度高于其他模型。