Quantification method of pore structure in low illuminance borehole images based on pixel spatial information
Given the technical challenges of insufficient image clarity and inaccurate quantitative characterization in the in-situ detection of pore structures on low reflection rock walls,this paper proposes a quantification method of pore structure in low illuminance borehole images based on pixel spatial Information.By synchronously utilizing borehole wall images and point cloud data to obtain pixel spatial feature information of non-standard cylindrical borehole shapes,the quantification process of pore structure in low reflection characteristic rock layers under complex geological conditions is achieved.Firstly,based on the low illumination borehole wall image features with alternating light and dark textures that are often formed in the actual drilling environment and testing process,a borehole wall eccentricity image correction model that is suitable for the actual hole testing environment is constructed to form a cosine light and dark texture suppression function that can effectively weaken the hole wall light and dark texture phenomenon.Subsequently,a low illumination borehole wall image enhancement algorithm based on detail feature weighted fusion is proposed to enhance the texture information of low illumination borehole wall images.Finally,combining the division of pixel spatial cells and the calculation of horizontal and vertical scales of pixel spatial points,a pore structure quantification method utilizing pixel spatial information is formed.At the same time,combined with practical case analysis,the correctness and superiority of the method proposed in this paper are verified.The results show that the method can obtain pixel spatial feature information of borehole walls in non-standard cylindrical drilling shapes,which can provide a new technical method and means for in-situ detection of pore structures in low reflection characteristic rock layers under complex geological conditions.
rock mechanicspore structureimage processingquantitative characterizationpixel spacelow illumination images
NETL
NSTL
万方数据
