Recovery of geometric properties of binary objects from blurred images

Geometric properties such as area and perimeter are commonly used in the analysis of binary objects. Images of binary objects often appear blurred due to the imaging process, and one needs to use thresholding techniques to convert the image to a binary image before such geometric properties can be computed. In this paper, we show that by using fuzzy definitions of geometric properties, one can make reasonably accurate measurements of geometric properties directly from blurred images, and the results so obtained are generally superior to those obtained by thresholding