A method for detecting and removing scrolling texts comprising a step of using an adaptive transient difference processing of video communication to conduct frame calculation, wherein the adaptive transient difference processing takes first N frames fjkt−N and a current frame fjkt, and subtracts them to obtain a frame difference; and if the frame difference is greater than a threshold value, it is determined that the current frame fjkt has scrolling texts; and interpolates the first N frames before the current position of the scrolling texts to replace the current frame fjkt to achieve the goal of hiding the scrolling texts during video communication to enhance the viewing effect.