Nonlocal Mean Filtering Algorithm for Low Contrast Images and its Application

Main Article Content

Tianyi Guan


High background noise, cracks, fuzzy boundaries image containing the chromatism, etc are the common problems faced in the low contrast image recognition, this paper takes the core fracture identification of two-dimensional section as an example, and highlight the point, simplify the problem, this paper only considering the three dimensional images of two-dimensional cross section along the direction perpendicular to the core shaft, focusing on the identification of disc core cracks within the dark grey. Each voxel based on 3D digital images corresponds to a gray value. The smaller the value, the blacker the corresponding voxel will be. The larger the value, the whiter the corresponding voxel is. By fine-tuning the background color difference, filtering and denoising, marking the non-crack area, secondary denoising by graphics method and other algorithm methods, the identification efficiency of the crack area in the two-dimensional cross-section diagram of the core column is effectively improved. This method can also be used as a solution to other problems in similar scenes.

Article Details

How to Cite
Guan, T. (2021). Nonlocal Mean Filtering Algorithm for Low Contrast Images and its Application. CONVERTER, 51 - 59.