Line Moments and Invariants for Real Time Processing of Vectorized Contour Data

In this article a new concept for real time analysis of images based on the vectorized contours is presented. Given the polygone approximation of contours the geometric moments can be formulated for line contours. The resulting line moments (LM's) can be calculated for an arbitrary collection of contour fragments. Neither the existence of an area nor a closed contour is prerequisite to characterize an object by line moments. Due to these properties of the LM's time consuming preprocessing of noisy object contours is not required. Even objects with structured surfaces can be characterized by LM's. The formalism to apply the geometric moments to contour patterns is derived. A direct and a recursive algorithm to efficiently compute the LM's is given. As an application of the LM's an analysis system for structured textures is presented and results are discussed.