The CC ShareAlike terms merely say that whenever you do make a derivative work of the image, distribution of that derivative work must be under ShareAlike (or CC-approved ShareAlike-compatible) terms. The relevant unsettled question is exactly when a derivative work is created, which lies in the domain of copyright law, not the license terms.
Certainly modifications to the image itself will need to be under ShareAlike terms, but the degree to which that requirement extends to a surrounding context is slightly less clear. There is some relevant case law (from the United States):
The U.S. case Drauglis v. Kappa Map Group, LLC found that the use of a CC BY-SA image as cover art for an atlas did not extend the ShareAlike requirement to the whole atlas. The court found use of the image as cover art to be a compilation of separate works rather than a single derivative. This is not exactly the same as inline inclusion in a blog post, but it may be the closest case law that currently exists.
Since you are considering an online blog specifically, there may be additional factors in favor of the image being a completely separate work. If the blog post is written in HTML and does not directly include the image (e.g., as a base-64 embedded data blob) but instead refers to it by URL, the U.S. case Perfect 10 v. Amazon found that referring to an image resource by URL is not equivalent to displaying the image. The fact that a web browser will attempt to render whatever resource is found at that URL does not mean the author of the HTML page is responsible for displaying that image.
Specifically, the Perfect 10 ruling says:
Providing these HTML instructions [i.e., to display someone's copyrighted image] is not equivalent to showing a copy. First, the HTML instructions are lines of text, not a photographic image. Second, HTML instructions do not themselves cause infringing images to appear on the user’s computer screen.
From such a legal conclusion, I think there's a strong argument that referring to an image by URL in an HTML page would not inherently create a derivative of the image, either, since the HTML instructions do not include the image itself.