In the evaluation of tear film stability based on corneal topography, a pretreatment algorithm for tear film video was proposed for eye movement, eyelash reflection and background interference. First, Sobel operator was used to detect the blur image. Next, the target image with highlighted ring pattern was obtained by the morphological open operation performed on the grayscale image. Then the ring pattern frequency of the target image was extracted through the Hough circle detection and fast Fourier transform, and a band-pass filter was applied to the target image according to the ring pattern frequency. Finally, binarization and morphological closed operation were used for the localization of the ring pattern. Ten tear film videos were randomly selected from the database and processed frame by frame through the above algorithm. The experimental results showed that the proposed algorithm was effective in removing the invalid images in the video sequence and positioning the ring pattern, which laid a foundation for the subsequent evaluation of tear film stability.
As the most important refraction part in the optical system, cornea possesses characteristics which are important parameters in ophthalmology clinical surgery. During the measurement of the cornea in our study, we acquired the corneal data of Orbscan Ⅱ corneal topographer in real time using the Hook technology under Windows, and then took the data into the corneal analysis software. We then further analyzed and calculated the data to obtain individual Q-value of overall corneal 360 semi-meridian. The corneal analysis software took Visual C++6.0 as development environment, used OpenGL graphics technology to draw three-dimensional individual corneal morphological map and the distribution curve of the Q-value, and achieved real-time corneal data query. It could be concluded that the analysis would further extend the function of the corneal topography system, and provide a solid foundation for the further study of automatic screening of corneal diseases.
