{
"format": "cvmat",
"width": width,
"height": height,
"image": <blob>
}
The current tuples generated from video sources look ilke this. In raspivideo, it also wanted to output raw images and JPEGs. So, I defined the format like below:
{
"format": "raw",
"color_model": "rgb",
"width": width,
"height": height,
"image": <blob>
}
Instead of "cvmat", I defined "raw" format containing raw pixel values in a layout specified by the "color_model" field. The "color_model" field is only provided when the format is "raw". "raw" + "bgr" is equivalent to the current "cvmat" format.
I think this is more general and flexible than using the term "cvmat". What do you think?
The current tuples generated from video sources look ilke this. In raspivideo, it also wanted to output raw images and JPEGs. So, I defined the format like below:
Instead of "cvmat", I defined "raw" format containing raw pixel values in a layout specified by the "color_model" field. The "color_model" field is only provided when the format is "raw". "raw" + "bgr" is equivalent to the current "cvmat" format.
I think this is more general and flexible than using the term "cvmat". What do you think?