The above is only an example to illustrate the value range of N, which may be limited in the same round of the game, and is limited according to the game time. The value range of N is not limited in the present disclosure.

Assuming that the first image frame from the left side view of the game desktop is image a, the synchronous image synchronously acquired at the top view with image a is image b, and the target object can also be detected in at least one image frame in the N image frames before image b, it can be considered that the specific target object is occluded during a period of time before image a is acquired from the left side view of the game desktop, resulting in that the earliest time point, when the target object appears, determined in the images acquired from the left side view of the game desktop in the present round of the game is incorrect. Thus, the game progress stage is switched at the earliest time point corresponding to the image frame where the target object is detected in the N image frames before image b in the images acquired from the top view.