Explanation of Terms

• $\star$ 2-D error - the distance, in the image plane, between the detected (real) image point and the model point that has been homographized and distorted by the computed model
• $\star$ 3-D error - the average distance, in three dimensions, between the detected (real) point and the ray projected by the camera toward the homographized and distorted model point
• $\star$ accuracy - the closeness of numerical results to actual data (see precision)
• $\star$ coordinate adjustment - actually moving the estimate of corner points that have been detected so that they make a nicer model of the homography
• $\star$ corner - a point in an image with a sharply defined image gradient (strong gradient value) (see interest point)
• $\star$ corner detection - searching images for corners and listing them in coordinated lists so that they can be analyzed
• $\star$ distortion - the movement of points from reality to an image due to an imperfect camera lens
• $\star$ extrapolation - interpretation of data for a result beyond the domain of previously obtained values
• $\star$ extrinsic parameters - parameters describing the relationships between the camera and the image planes
• $\star$ homography - the 3-D geometric relationship between two sets of points (in this case, points on planes)
• $\star$ image plane - the plane containing an image taken, not necessarily a simple plane like the model plane (see image plane)
• $\star$ interest point - any point in an image that is interesting for its image or gradient value (see corner)
• $\star$ intrinsic matrix - a matrix containing five parameters that describe the characteristics of a digital camera
• $\star$ intrinsic parameters - with reference to a digital camera, its focal length, skew and lens principal point
• $\star$ model plane - the plane containing the model points, measured by hand (in this case, the plane z = 1) (see image plane)
• $\star$ optimization - adjustment of a set of parameters for the minimization of a function of those parameters, called a "cost function"
• $\star$ parameter - any value that is changed by a function, such as an optimizer
• $\star$ precision - the ability to accurately reproduce results (see accuracy)
• $\star$ radial distortion - distortion in the radial direction (away from the center of the lens)
• $\star$ tangential distortion - distortion in the tangential direction (parallel to a line through the center of the lens)
• $\star$ transformation - a series of translations, rotations and scales defining a geometric relationship between two objects or points