Skip to Main Content
A method is described for the rapid reduction of intensity matrix image data to chain-coded descriptors. The method also yields some topographic and numerical information about regions in the image. In this method, the chain code is generated and stored in an array which is isomorphic to the image array. Computation other than that associated with the intensity classification (feature) operator is minimized by the use of decision tables for chain code manipulation. The method is capable of processing a 300×200 gray scale image in two seconds using a gradient and threshold operator on a 500K instruction per second processor.