Influence of uneven lighting of microscope on petrographic images and its calibration methods

The combination of digital image analysis and deep learning algorithm is essential for automatic determina⁃ tion of vitrinite reflectance and intelligent recognition of coal macerals. Generally speaking，the images are collected in a large field of view to obtain comprehensive information. Three optically isotropic standard samples，i.e.，sapphire，yt⁃ trium aluminum garnet (YAG) and gadolinium gallium garnet (GGG)，are selected to study the influence of uneven lighting of microscope on the characteristics of petrographic images，as the images of standard samples are collected in large view of field. Theoretical analyses are conducted as the illumination of petrographic microscope are point light and area light. Finally，the digital image processing technologies based on lowpass filter and top⁃hat transformation are used to calibrate the uneven⁃lighted images. The coal samples are used to verify the calibration effect. The distribution of gray level ( reflection intensity ) in sapphire， yttrium aluminum garnet ( YAG ) and gadolinium gallium garnet (GGG) are similar to concentric circles as the center point is located in the center of the image. The gray val⁃ ue decreases from the center to the edge of the image. The gradient of gray value in the center of the image decreases slightly，while that of the point far away from the center decreases dramatically. The distribution of gray value in the standard samples changed with the optical axis of the microscope. The uneven lighting is sure to occur whether the light source of the petrographic microscope is point light or area light according to the theoretical analysis. The reflec⁃ ted light on the surface of blocks is captured by the CCD camera and some shadows will be formed in the digital im⁃ age. The shadow image caused by uneven lighting can be calibrated by lowpass filter and top⁃hat transform. The mean gray value of the calibrated standard image is greatly improved while the standard deviation is significantly reduced. The gap between the first and third quartile is within 2 gray value. The standard deviation and uniformity of the gray value in the standard images after top⁃hat transform are lower than those of lowpass filter，while the resolu⁃ tion capability of pixel difference after top⁃hat transform is better than that of lowpass filter. The petrographic ima⁃ ges can be calibrated by both lowpass filter and top⁃hat transform as shown by the results of single coal and blending coal. The top-hat transform is preferred in petrographic image calibration.