The presence of fractures significantly influences the physical and mechanical properties of rock masses. However

automated extraction of fractures from images of rock mass exposures frequently encounters challenges such as incomplete results

low signal-to-noise ratio

reliance on empirical parameters

and poor robustness. In response to these issues

a fully automated optimization procedure for the extraction of fractures is proposed. Firstly

the introduction of a generalized gamma correction is employed to preliminarily enhance the contrast between fractures and rock wall surfaces. Subsequently

by considering the mutual influence among pixels along fracture paths

a grayscale transmission algorithm is devised to improve the continuity of fractures. Finally

an improved Frangi filter is utilized for fracture extraction while effectively suppressing the response of noise pixels. The results indicate that the proposed procedure cleverly integrates two major characteristics of fractures: low grayscale and high linearity

and the coordination between the stages of image enhancement and fracture extraction significantly ameliorating the issue of uneven fracture contrast. Furthermore

while ensuring complete fracture extraction

it efficiently prevents the generation of noise and pseudo-fractures. The procedure demonstrates high robustness across various images of rock mass exposures. Comparative analysis with commonly used fracture identification algorithms highlights the advantages of the proposed procedure.