Error model for a three-dimensional machine vision system

The intrinsic measurement error of a dual video camera-based three-dimensional (3-D) machine vision system is analyzed. The sources of error are assumed to stem from common arrangement inaccuracies and limited precisions of the imaging device. The resultant error equation for a point in 3-D space is related to system parameters such as pixel resolution, camera focal length, camera position, separation and orientation. For normal camera to object distances, the accuracy of the vision system is strongly influenced by camera orientation errors and least affected by limited resolution of the imaging device.