【摘要】 针对地震勘探数据采集工作中存在大量废道的问题,提出了一种基于BP神经网络的废道自动切除方法。通过对废道的产生机理和干扰特性的深入分析,提取能够反映废道本质特征的参数,结合BP神经网络强大的模式识别和自主适应能力,设计了一个三层的BP神经网络结构;最后,采用实际地震资料对处理算法进行验证。结果表明,该方法废道处理的准确率高、处理速度快,并且具有良好的适应性,能够满足地震勘探工作对废道处理的要求。更多还原

【Abstract】 According to a large amount of dead traces in the seismic exploration data,this paper puts forward a dead traces automatic resection method. With the deep analysis on the interference mechanism and characteristics of dead traces,we extract the parameters which can reflect the essential characteristics of dead trace.Combining with pattern recognition and adaptive ability of BP neural network,we design a three layers BP neural network structure.Finally,we validate the processing algorithms using the actual seismic data.Results show that thedead tracesprocessing method has high accuracy,high recognition rate,and has a good adaptability.It can meet the requirements of seismic prospecting work on dead traces processing.更多还原