Calibration Tools Overview

A calibration tool creates calibration objects (transforms) that relate locations in image space to locations in physical space. You typically use these calibration objects as the client coordinate transform in pel buffers but they can be used for other purposes also, such as the transform that defines the warping for the Image Warp tool.

CVL provides two different and distinct calibration tools: the grid-of-dots calibration and the feature correspondence calibration. The grid-of-dots calibration is the older tool that uses an image of a grid-of-dots calibration plate to compute a linear transform (type cc2XformLinear) or a nonlinear polynomial transform (type cc2XformPoly). This tool has one class that does everything from the feature extraction to transform calculation and uses only grid-of-dots calibration plates.

The newer Feature Correspondence Calibration tool creates a nonlinear transform (type cc2XformCalib2) that models radial optical distortion and perspective distortion. The tool has two components: feature extraction and calibration. The feature extractor is specifically designed to extract features in an image of a checkerboard calibration plate The calibration component can create a calibration transform using the output of the feature extractor. This two step design allows you to create a cc2XformCalib2 calibration object from your own set of features if you choose not to use the Cognex checkerboard feature extractor. For example, you can use any set of corresponding image points and physical points to create a cc2XformCalib2 calibration object using this calibration tool.

The cc2XformCalib2 calibration object has been found to be more accurate than the grid-of-dots calibration objects for various distortions such as perspective distortion.

The grid-of-dots calibration and is discussed in the following section and the feature correspondence calibration and is covered in Feature Correspondence Calibration.