ANALYSIS OF GRAIN SIZE MEASUREMENT METHODS IN SEMIAUTOMATIC IMAGE ANALYSIS SETUP

摘要

This research focuses on the analysis of grain size measurement methods according to ASTM standard (ASTM E1382) in a semi-automatic image analysis setup. The grain size measurement methods are implemented on a software tool and verified using images with known grain size numbers. The software tool developed for this purpose is able to perform all the measurement methods specified by the standard, namely intersection count, grain count, grain area, chord length and grain boundary length per area methods. Image processing techniques are provided within the software so that the user can adjust the quality of the photomicrographs before calculating the ASTM grain size number as well as other relevant statistical parameters. The accuracy inherent to each measurement method is investigated. Moreover, the effect of sampling techniques are also studied. From a limited set of images used, it is found that, on the overall, the grain count method has provided the most accurate result. The minimum sample size of 50 grains affects the measurement accuracy more than the sampled locations. Each measurement method in the software can provide the ASTM grain size number within 15% of that using the manual method. It is therefore expected that the software can be used to shorten the prototyping and pre-production time required. Additionally, it can be used periodically in production as a quality control measure.