In the process of measuring the diameter of the air film holes in turbine blades, visual technology is widely used. However, the selection of a clearness evaluation method and clearness evaluation operator for the pixels in the images obtained by photography is a crucial problem. In view of the small diameter of the air film holes in turbine blades, strong reflective characteristics of the material, difficulties in obtaining a focused image of the oblique holes, and difficulties in edge extraction, a five-axis measurement system was built using a three-coordinate measuring machine and a two-axis rotary table to take a series of pictures of the air film holes in turbine blades. In terms of image processing, different clearness evaluation functions were used to test images at different focal positions, including Sobel function, Brenner function, Tenengrad function, Laplacian function, etc. By comparing and analyzing the clearness evaluation curves obtained by these functions, the most suitable clearness evaluation function was selected. In order to make the results more convincing, the selected function was normalized to ensure its evaluation results were comparable across different images. Finally, a suitable clearness evaluation function for the air film holes in turbine blades was derived and its effectiveness was verified. This achievement lays a foundation for the improvement of the Sobel operator in the future, which is helpful to improve the accuracy and stability of measuring the diameter of air film holes in turbine blades.