Template matching for videos

I’m currently planning out an application that takes a video input, searches for a pre-defined event (an existing image) in said video and then triggers actions upon locating a defined event. The actions I am currently looking at are to take the frame where the event occurred, save it as an image and call an API.

I’ve been looking at OpenCV’s template matching to identify the visual event, but so far I have mostly found resources that do this for only images or real-time video feeds. As the video input could be 1-3 hours long, real-time search is not a preferred solution.

Is it possible to use template matching as a solution for this application or are there better-suited alternatives?

just read the video.

why are you worrying about something you haven’t encountered and won’t encounter?