基于改进人工蜂群算法与随掘超前探测的地震去噪技术研究

Research on seismic denoising technology based on improved artificial bee colony algorithm and advance detection with excavation

针对随掘地震信号中噪声的分离问题,研究提出了一种结合改进的人工蜂群算法和独立分量分析的地震去噪模型。该模型通过分组引导结构增加算法的多样性,并引入比例因子与程度系数增加模型的泛化能力。在实验结果中,模型在Sphere函数上的最佳函数值从5000次评估的7降至15万次评估的-205,在Quartic函数上的最佳函数值收敛于-6;经改进人工蜂群去噪后的信号X分量、Y分量以及Z分量,其中分量信号的能量特征超前探测的震源特征相符合,表明经分离后的信号与震源信号相接近。研究提出的模型能更有效地分离和去除地震信号中的噪声,特别是在处理复杂信号和提高噪声抑制能力方面显示了更加优越的性能。结果表明,研究方法在复杂信号的处理和噪声抑制方面具有明显优势,通过引入分组引导式搜索策略和动态更新比例因子及程度系数的策略,以此方式提升算法在面临复杂地质场景时的适应性和优化性能,为信号去噪带来了新的研究思路。

Aiming at the problem of separation of noise in seismic signal with excavation,a seismic denoising model combining improved artificial bee colony algorithm and independent component analysis was proposed.The model increases the diversity of algorithms by grouping guidance structure,and introduces scale factors and degree coefficients to increase the generalization ability of the model.In the experimental results,the optimal function value of the model on Sphere function decreases from 7 of 5000 evaluations to-205 of 150000 evaluations,and the optimal function value on Quartic function converges to-6.The X component,Y component and Z component of the signal after the improved artificial bee colony denoising,in which the energy characteristics of the component signal are consistent with the source characteristics of the advance detection,indicating that the separated signal is close to the source signal.The proposed research model can separate and remove noise from seismic signals more effectively,especially demonstrating superior performance in processing complex signals and improving noise suppression capabilities.The results indicate that the research method has significant advantages in processing complex signals and noise suppression.By introducing a group guided search strategy and a strategy of dynamically updating scale factors and degree coefficients,the adaptability and optimization performance of the algorithm in complex geological scenes are improved,bringing new research ideas for signal denoising.