A methodology to measure the volume of spheroid and oblong solid bodies based on artificial vision technique

Abstract

A methodology for assessing the volume of spheroid and oblong solid bodies is presented. Samples were mounted on a revolving platform that was driven by a computer-controlled stepping motor. Four hundred views (photographs) of each sample were acquired as they were uniformly rotated in the azimuth direction. The image processing was based on the artificial vision technique called segmentation. Using the information of the instantaneous radius and the small angle of rotation in each step, the numerical integration of the volume was performed. Images were captured using a CCD camera and the entire system was controlled by a routine developed in LabVIEW$^{\mbox{\texttrademark }}$ 6.1. Two sets of geometrical bodies (polystyrene cylinders and spheres) and three kinds of biological samples were measured. For the sake of comparison, each body was also measured by means of both a micrometric caliper and the displaced volume of water inside a vessel. The ANOVA correlation parameters between the proposed methodology and the hydrostatic procedure were found to be: r = 0.9924 and p = 0.0001, with $\alpha $ = 0.05. The coincidence between the results obtained with artificial vision and the hydrostatic technique was greater than 98% for spheres and cylinders. On the other hand, it was only up to 95% for the samples with non-regular shaped bodies (chicken hearts, kidneys and carrots). The purpose of the paper is to discuss in detail a simple technique which could be of interest to students of science and engineering.