It's a good point, you can't, you can only say it's close enough to perpendicular to be accurate. But when the resulting difference is a millimeter, that margin of error becomes significant.
The trouble is that literally every part of VAR ultimately ends up being somewhat arbitrary.
Which frame do you pick to draw the lines on? You need a camera that has a high frame range, then cycle to the exact point where the ball leaves the foot, then go one frame back. However even this is hard to determine, as on a 2d image, if the foot is overlapping the ball then there's no way to know if it's still touching the ball or not.
How high is the resolution of the image? Lower resolutions will cause blur, which may enlargen areas of a players body.
How is the focus of the camera set? Even a small difference will cause players in the background to appear larger than they actually are.
Is the camera aligned with the pitch properly. Since your using a 2d image from that camera to determine the result, the camera needs to be perfectly horizontal in order for the perspective to be judged accurately, as you mention with the perpendicular lines.
All these variables means it's actually impossible to have a 100% objective decision. If the player is a yard over, then fine. If it's a millimeter, there's a good chance that it's incorrect.