Influence of the machine vision lens to the accuracy in contactless 2D-metrology

摘要

A model for quick estimation of the accuracy of the contactless 2D-measurement is developed. Perspective errors of entocentric lenses and distortion errors of entocentric and telecentric lenses are respected. The perspective error starts from the standard deviation of the working distance. The distortion contribution starts from the often publicized maximal distortion value in%. The distortion contribution is divided in two parts: The size part considers the variation from calibration width and to width to measure, the field part considers the position variation of the piece under test in the field of view. Two graphs show typical and generally valuable functions. The superimpose of different contributions is realized as sum of quadrates. Due to this, the result can be interpreted as the standard deviation of a contactless measured width. All results have the dimension pixel. Three examples are presented. The first one shows the accuracy limits in a typical machine vision application using different entocentric lenses. The second one discusses an entocentric lens measuring different object sizes. The third one uses a telecentric lens at same measuring condition. Here, the estimated accuracy is ten times higher.