Enhancement and evaluation of the geophysical inversion model based on the human visual system

摘要

Geophysical inversion algorithms often introduce artifacts near sharp intensity transitions and image boundary. We presented a iterative restoration algorithms to restore the true inversion image and remove the artifacts. We also presented a ringing metric to evaluate the quality of images restored based on the human visual system. According to types of ringing artifacts, two ringing metrics is used to assessment the inversion image based on the analysis of ringing artifacts. Finally, an overall ringing metric is presented based on the two ringing metrics. The experimental results validate the proposed restoration and evaluation methods perform well.