The application of an image processing technique is presented for the volume estimation of very irregular small biomaterials (wheat and rice-paddy grains). Two common cylindrical small biomaterials, the Alvand variety of wheat grain and the Neda variety of paddy grain were considered for examination. The captured images were exported to be processed by an image processing software (ImageJ) and the edge-extracted image was used in SolidWorks for the 3D reconstruction of the model. The revolved images in the SolidWork were used to estimate the volume of the examined grains. The estimated volume was then compared with the conventional mathematical expression and also with the real volume measurement using the fluid displacement method. Volume estimation using machine vision and image processing techniques has a considerably lower mean error (9.5%) in comparison to the mathematical error (14.7%). The average value of cylindricity for Alvand wheat was found to be equal to 82.34% at a moisture content of 11.83%. The new cylindricity factor had a significantly smaller standard deviation in comparison to the standard deviation of the sphericity factor for the examined cylindrical crops (61.5% for the wheat grains and 59.6% for the paddy grains). The new cylindricity factor can be used for the heat and mass transfer modelling of cylindrical crops.