针对北斗观测数据中存在的小周跳难以探测的问题,提出了基于SVD分量包络的周跳探测方法。首先利用伪距和相位相结合的方法构造出周跳检验量,然后对周跳检验量构建Hankel矩阵进行奇异值分解,得到包含突变信息的子空间。再对子空间逆变换得到的分量信号进行包络分析。实验结果表明,利用奇异值处理去噪后,进行包络变换能正确探测出小周跳。 Aiming at the problem that the small cycle jump is difficult to detect in Beidou observation data, a new cycle jump detection method based on SVD component envelope is proposed. Firstly, we use the combination of pseudorange and phase to construct the cycle-slip test, and then construct the Hankel matrix for the cycle-slip test to carry out the singular value decomposition to obtain the subspace containing the mutation information. Then the component signals obtained by the inverse transformation of the subspace are enveloping analysis. Experimental results show that using the singular value processing to denoise, the envelope transform can correctly detect the small cycle jumps.