Grain size is one of the crucial parameters in the microstructure analysis of high strength aluminum alloy. This information is commonly derived based on manual processes. However, these manual processes may take long time and are error prone. Nowadays，the rapid development of the digital image processing and the pattern recognition technologies provides a new methodology for the quantitative metallographic analysis. Artificial intelligence utilized in realizing automatic metallographic analysis can overcome the drawbacks of the manual processes. Through this paper we present a new method of digital image processing for determining the grain size of the metallographic images. To derive the grain size of the digital metallographic images, the digital image processing is applied to extract grain boundary by proposing a new edge detection algorithm based on the fuzzy logic. Extensive metallographic images with different qualities were tested to validate this method. Practical application use cases are presented here. The grain size is calculated in accordance with American Society for Testing Material (ASTM) standards.