论文部分内容阅读
用常规反褶积压制周期长于100毫秒的多次波时,有两种系统误差。一种误差是时窗效应,它是由根据数据的有限部分计算实际自相关函数而引起的。另一种误差是周期性假设所引起的,这种假设只有在一维介质、零偏条件下才能成立。误差的幅度随多次波周期长度的增加而增加。本文描述并证实了一种能克服常规一维法中两种系统误差的预测反褶积方程的精确二维解。该法可用于简单波场系统或捕波系统以及复杂或微屈多次波中。该方法并不要求将时窗设计成算子大小的6～10倍,在较大的传播角内均很精确。从经典预测意义上看,该法所用的公式是严密的。
There are two systematic errors when conventional deconvolution is used to suppress multiples with periods longer than 100 milliseconds. One type of error is the time window effect, which is caused by calculating the actual autocorrelation function based on a finite portion of the data. Another kind of error is caused by the periodic hypothesis, which can only be established under the condition of one-dimensional medium and zero bias. The magnitude of the error increases as the length of the multiple cycles increases. This paper describes and demonstrates an exact two-dimensional solution to the predictive deconvolution equation that overcomes both systematic errors in conventional one-dimensional methods. The method can be used for simple wavefield systems or acquisition systems, as well as complex or flexion multiples. The method does not require that the time window be designed to be 6 to 10 times the size of the operator, and that it is accurate over a large propagation angle. From the classical predictive point of view, the formula used by the law is tight.