基于FCM聚类约束的直流电阻率法与地震走时成像法二维联合反演

通过引入模糊均值聚类(FCM)模型约束函数对电阻率与速度进行约束,开展二维直流电阻率法与地震初至波走时成像法联合反演研究.在地下浅层结构勘探中,通常低电阻率的地质体具有低速特征,较高电阻率的地质体表现为较高的地震波速度.直流电阻率法因为低电阻率区域吸引电流而对其敏感,地震走时成像法因为射线集中在高波速区而对高速体敏感,因此,两者联合成像能够大幅度提高反演效果.合成数据反演表明,直流电阻率法和地震初至波走时联合反演对于两类地质体的分辨能力均有提升,能够优势互补.尤其是引入FCM模型约束进行联合反演,根据已知物性进行监督学习,进一步提高了反演质量,改善了成像模型的分辨率.

Two-dimensional joint inversion of DC resistivity method and seismic traveltime tomography method based on the FCM cluster constraint

Two-dimensional joint inversion of Direct Current resistivity method (DC) and first-arrival traveltime seismic tomography method have been studied in our paper, by adding the Fuzzy C-means clustering model (FCM) constraint function to inversion cost function in order to constrain the resistivity and velocity. In the near surface geophysical exploration, conductivity geological bodies always have low seismic speed, whereas resistivity geological bodies usually have high seismic wave speed. The DC resistivity method is sensitive to conductivity bodies because this region attracts current, and seismic traveltime tomography method is sensitive to high-speed bodies because the rays are concentrated in the high-speed region. Therefore, the combination of two data sets can greatly improve the quality of inversion result. The synthetic examples showed that the joint inversion of DC resistivity method and first-arrival traveltime seismic tomography method improves the resolution ability for both resistivity and speed geological model, and complements each other. Particularly the joint inversion with FCM model constraint, in which supervised learning is carried out by using priori physical properties, further improves the quality of inverted models and the resolution of geological units.