Qu, Shan (China University of Petroleum) | Zhou, Hui (China University of Petroleum) | Chen, Hanming (China University of Petroleum) | Zu, Shaohuan (China University of Petroleum) | Zhi, Ling (China University of Petroleum)
In this abstract, based on the blending matrix expression of different encoding methods deduced by using Principle of Stationary Phase, two separation methods are presented to separate the blended vibroseis data. In addition, before separation, we apply deconvolution to the blended vibroseis data as a preprocessing tool. The blending problem can be solved by an iterative method, which is extended from an iterative method proposed by Doulgeris (2010). A new pseudo-deblending matrix is defined to make it suitable to all encoding methods. The blending problem can also be treated as a regularization inversion problem with a sparse constraint after deconvolution. Both methods are effective and they have strong applicability to different encoding types. Deconvolution as a preprocessing tool is also very effective. Several numerical examples are given to show the effectiveness of the deconvolution preprocess and the two methods for different encoding types.