The last step in the loop is an error check. This routine calculates the errors, on the plane in which we have been working, between the points measured in Sec. 3.1 and the homographized points from the model file. The goal of this project is to have an accuracy such that the maximum error in 2-D is well under one pixel. If the result is not close to the desired value, the process continues with Sec. 3.1. At the moment, the maximum 2-D error is the number being used to decide how good a model is, rather than the average error.