Once you find or create the kernel of a pin—an enemy piece lined up with its king with nothing between them—you have to (a) figure out what piece you can use to impose the pin and (b) get it onto the square it needs. Sometimes the path to the pinning square is not clear, and in that case there are standard methods to consider to remove the obstructions.
We start with the easy case to the left. Searching from the Black king outward, you see the knight on e4 is ripe to be pinned. It’s on a diagonal, so the pin would require a queen or light-squared bishop. The bishop is available but its path to the pinning square is blocked by White’s own rook on e2. When one of your own pieces obstructs the path of another, the usual solution is to vacate the obstructing piece in a violent manner that requires a reply. Here the rook can take the knight on the pinnable square, so White plays RxN; Black replies RxR; and now the way is clear for White to play Bd3, pinning the rook and taking it a move later. White ends up a knight to the good.
By the way, this is another case where White also starts with a relative pin in place: if Black moves his knight, White has RxR. Though our full study of relative pins will come later, we can ask briefly whether it makes sense for White to exploit this one by attacking the knight with Bg2. Black has a natural response: he saves his knight and uses it to protect the rook by playing Nd6. Now White plays RxR; Black plays NxR—and White has Bxb7, for notice that Bg2 also cross-pinned Black’s knight to the loose pawn on b7. But this only nets a pawn. White still should win, but not as easily as he would with a whole extra piece.