Implementation of the morphological hit- or-miss transform using an acousto-optic architecture

Abstract

In this paper we describe an acousto-optic morphological image processor used to implement a morphological hit-or-miss transform. Morphology is an image algebra technique whereby the shape and structure of an input object is transformed by a smaller object referred to as a structuring element. The two most fundamental morphological operations, erosion and dilation, can be implemented as a convolution followed by a threshold. We perform the convolution optically and the threshold digitally. The input to the system is displayed on an Epson 1020 LCTV, and the system utilizes an acousto-optic Bragg cell in the Fourier plane of a coherent optical correlator to generate the desired structuring element. The hit-or-miss transform is the intersection of two eroded images. One image is the input object eroded by the desired structuring element, and the second image is the erosion of the complement of the input by the complement of the structuring element. This transform is useful for extracting specified features from images, as is necessary in automatic target recognition. Experimental results are presented.