Abstract <p>The scientific novelty of this work lies in the application of known methods (the Feret diameter method, the equivalent diameter method) in a new field (design and construction of elements in microelectronics). Accurate measurements of grain sizes of metal films and voids in TSV structures are critical to improving the reliability and performance of micro- and nanoelectronic devices. Manual methods of analyzing the morphological characteristics of materials require significant time and are subject to subjective errors. This paper presents an automated method for calculating grain sizes based on the processing of images obtained using a scanning electron microscope. The methodology uses two approaches to calculate the average grain size: the Feret diameter method and the equivalent circle method. The correlation between the results of these methods confirms the correctness of the segmentation and the high accuracy of the analysis. Experimental studies have shown that the proposed methodology allows the effective detection of grains and voids even in images with low contrast and high noise levels. The obtained results demonstrate the versatility of the method, its accuracy, and its reproducibility, as well as the possibility of integration into quality control processes and the design of microelectronic systems. Automation of analysis significantly reduces the influence of the human factor, shortens data processing time, and presents new opportunities for optimizing the production processes of micro- and nanoelectronic devices.</p>