Dear valued lainons. This summer I am embarking on a project to turn 30+ hours of video footage into clean, tidy data. Unfortunately, I have absolutely no knowledge of anything computer vision related, so any help to set me off in the right direction would be greatly appreciated. I am semi-proficient in Python and proficient in R. The project basically entails watching a scoreboard of a turn-based game, and transcribing all relevant data for each turn into a tidy dataset. The scoreboard and turn counters are all static for the most part. I might tab out here and there, or shift them around a little, but I can hopefully resolve all that manually.
There are 6 players and let's say 400 turns. So, I want my data to look like picrel, where there would be 2400 rows total. That way, I can group the data both by player and by turn, and get a full summary of everything. Some examples of things I'll be looking for per turn:
>A player's score
>How long they took to end their turn (their name color changes once they end their turn. This is actually the entire point of the project, so this column is very important)
>How many times they ended and restarted their turn (how many times their name color changed)]
>and a bunch of other dogshit which I don't really know yet
What I want from (You): How would you go about doing this? Any guiding advice would be amazing before I dedicate days of my summer to bashing my head against this wall. If any of you have a good youtube series for me to follow, that would be fuarrrking amazing. Or perhaps a recommendation of language / program, because I honestly have zero knowledge on the matter. I know I'll have to use machine learning for this, and that's intimidating the fuarrrk outta me. So I figured I'd ask here before I start doing serious background research, which I will probably do right after this post. Much love.