<p>To maintain traffic flow and ensure safety, road administrators and transportation service managers in snowy regions require accurate visibility information. Image-based visibility evaluation indices, which reflect image contrast, have been utilized as a novel tool for this purpose. However, image contrast can appear similar even when actual visibility conditions differ. In such cases, perceived visibility can vary due to differences in contrast sensitivity. The luminance adaptation effect refers to the phenomenon in which the minimum perceptible brightness difference increases as the background luminance increases. This effect can significantly influence contrast sensitivity, particularly at nighttime. Therefore, introducing an index that accounts for contrast sensitivity changes due to luminance adaptation is expected to improve the accuracy of nighttime visibility estimation. The proposed method estimates visibility using two indices: one representing image contrast and the other representing changes in contrast sensitivity due to luminance adaptation. It is designed for use in edge computing environments, where computational efficiency is critical. To minimize computational cost, the contrast sensitivity-based index is calculated using DCT coefficients obtained during the image compression process. In summary, the proposed method enables visibility estimation that considers both image contrast and contrast sensitivity change, with minimal additional computational burden. The method was applied to real-world in-vehicle camera images, and its effectiveness was verified through experiments.</p>