One of the cool feature of Semantic Kernel is the ability to shield the programmer from the underling interface of the API you are using. When it is time to include images in the prompt, to use multimodality of the model, Semantic Kernel make it possible to perform this task in very few lines of code.
Being able to call a LLM with images is one of the key point in RAG system where you have documents with lots of images.
As usual you can find the code here: https://github.com/alkampfergit/ai-notebooks/blob/master/semantic-kernel-base/210-send-image-functioncall-new.dib