Air-void characteristics are necessary to provide concrete with resistance to freeze-thaw damage. These properties are conventionally determined according to the standard ASTM C457 procedures involving manually measuring the characteristics of air-void, aggregate, and paste regions on a polished concrete section under a microscope using a point or line traversing technique. However, ASTM C457 is time-intensive and operator-dependent, limiting its routine application. In this paper, a new digital image-processing technique is proposed for accurate and quick measurements of concrete air-void properties. To allow for a clear differentiation between cement paste, aggregates, and air voids, the proposed methodology relies on macroscopic photographs of polished concrete surfaces taken using a microscope, video camera, and image-analyzing software. A software called Image 5 was used in this study. A dedicated image-analysis algorithm was developed for automatic generation of quantitative air-void data by mimicking the ASTM C457 count point and linear traverse methodology. Good agreement between the new and traditional methods was confirmed in statistical analysis. Moreover, the effect of additives, including fly ash and air-entraining admixture, on air-void characteristics was investigated digitally. Digital image analysis indicated that the air void content was 4.3% in control concrete, reduced to 3.9% by fly ash, and increased to 5.7% by air-entraining admixture. The proposed method demonstrates high accuracy, repeatability, and suitability for commercial quality control and laboratory evaluation of air-void structure in modified concrete.