In this paper, the vibration process of the vocal folds of the examinee is sampled by laryngoscopy and used as sample data. Five major classes of filters are used to remove the background noise of the vocal fold image in separate cases to optimize the segmentation effect of the background and the vocal folds. In order to improve the extraction accuracy of vocal fold vibration contour points, the OTSU algorithm is used to set adaptive thresholds for different regions of the image, and binary segmentation is performed to segment the grayed-out vocal fold image and the background image. Based on the extracted vocal fold vibration data, we calculated and compared the normal and abnormal vocal fold vibration characteristics using modern popular vocal singing techniques. Inadequate mastery of modern popular vocal singing techniques will lead to 10 types of problems such as excessive guttural sound, which is reflected in the vibration characteristics of the vocal folds, such as complex and disordered waveforms, high and low frequencies, and reduced peaks, etc., and overall fails to show the beauty of popular vocal music.
