WHATS IN A SET OF POINTS

The problem of fitting a straight line to a planar set of points is re considered. A parameter space computational approach capable of fittin g one or more lines to a set of points is presented. The suggested alg orithm handles errors in both coordinates of the data points, even whe n the error variances vary between coordinates and among points and ca n be readily made robust to outliers. The algorithm is quite general a nd allows line fitting according to several useful optimality criteria to be performed within a single computational framework. It is observ ed that certain extensions of the Hough transform can be tuned to be e quivalent to well-known M estimators, thus allowing computationally ef ficient approximate M estimation.