A Cellular Automaton Processor for Line and Corner Detection in Gray-Scale Images

The design and VLSI implementation of a Cellular Automaton processor for the detection of lines and corners in gray-scale images is presented in this paper. The behavior of a number of different Cellular Automaton rules was investigated and it was found that certain rules result in transitions in the Cellular Automaton state-transition diagram that correspond to the masks required for the line and corner detection. More specifically, the one-dimensional Cellular Automaton of length 8, operating under rule 56 with periodic boundary conditions, is capable of generating different sets of mask operators for line detection, corner detection and dominant point detection (and, thus, for arbitrarily-shaped curve detection), depending only on the initial state of the Cellular Automaton, without any additional hardware cost for the implementation or the reconfiguration of different masks. The proposed architecture was designed and implemented on a single VLSI chip using 0.7 μm double-layer metal (DLM) CMOS technology. The behavior of the chip was successfully verified for all sets of masks for line detection, corner detection and dominant point detection.