A Study on Determination of Simple Objects Volume Using ZED Stereo Camera Based on 3D-Points and Segmentation Images

Abstract

Known visual information is needed to acquire exact pose of an object for controlling a picking system. There are several known visual information of objects such as depth, volume, mass, shape and boundary, etc. The volume and the depth of them are very important information for controlling the picking system. Until now, measuring 3D volume and depth rendering from one side of 2D projection using an image processing method are difficult. Previous researches using five cameras captured segmented image to estimate objects volume or concentrated only on 3D volume of objects around the principal axis. However, they did not estimate volume of several picking objects such as cubic and cylindrical shapes. To solve this problem, this paper present an image processing algorithm for the determining volume of simple objects using ZED stereo camera. In the proposed algorithm, 3D-points rendered from 2D projection combined with image segmentation are used to estimate the depth and surface areas of cubic and cylindrical objects, respectively. Experimental results show that object volume is compared with the conventional method measuring the object size by slide calipers.