If you use cv::cvtColor(gray, color, cv::COLOR_GRAY2BGR)
you should get a 3 channel image but there are no colors, because all the channels have the same value as the gray image.
The area of complete restoration of a Grayscale Image to RGB Image is yet under active reserach and mostly available in research papers, individual author documents etc. This has not yet been officially supported by OpenCV or OpenVINO yet. However, one can enhance the quality and extract meaningful information from a low resolution /blurred image/low contrast using OpenCV. This is possible through here - one can enhance the quality and extract meaningful information from a low resolution /blurred image/low contrast using image processing.
This has not yet been added to Depthai Examples yet, since it is a low power application. Let me know if this was helpful.