This study aims to improve the automation and accuracy of extracting parameters of rock mass structural planes. By considering rock mass characteristics of the structural plane in borehole images

a new method was developed for the pre-processing of structural plane image and the extraction of structural plane parameter. First

the noise in borehole images was removed with an adaptive median filtering method. An improved variance method was also proposed by comparing the gradient operator method and the maximum interclass variance method

which is more suitable for the actual borehole image segmentation. Then

a Canny operator was used to detect the edge of segmented images. Finally

the sine curve of the structural plane was obtained by the polynomial curve fitting

and thus the structural plane parameters including dip

dip angle and width were calculated. The comparison of results obtained by the proposed extraction method and the traditional manual recognition shows that the proposed method is more accurate. Therefore

it is demostrated that the proposed method is feasible and accurate

and its extraction efficiency is higher than the traditional manual method.