iteration

One random-sample-and-fit attempt in RANSAC's loop -- pick a minimal sample, fit, count agreement, then try again.

One iteration of RANSAC is a single pass through its loop: draw one random minimal sample, fit a candidate model to it, and count how many other points agree with that fit. RANSAC repeats this a fixed or adaptively chosen number of times, $k$ ; the RANSAC consensus guarantee (this section) is the formula for how the chance of finding a correct fit grows as $k$ grows.

Examples

Running RANSAC for $k=100$ iterations means trying 100 different random minimal samples and keeping the one with the largest inlier count.