allow IR16 format in k4a_transformation_depth_image_to_color_camera_custom()

[diablodale]

I request two lines of additional code be added to allow K4A_IMAGE_FORMAT_IR16 to be used as input into k4a_transformation_depth_image_to_color_camera_custom(). I see only benefit and no harm for the trivial change.

This is highly related to https://github.com/microsoft/Azure-Kinect-Sensor-SDK/issues/1188

Scenario

Transform the IR image from its native resolution into the color camera's resolution.

Problem Cause

k4a_transformation_depth_image_to_color_camera_custom() is artificially limited to only accept K4A_IMAGE_FORMAT_CUSTOM8 and K4A_IMAGE_FORMAT_CUSTOM16. Yet the entire workflow for it actually can support also K4A_IMAGE_FORMAT_IR16.

The artificial limitation is

https://github.com/microsoft/Azure-Kinect-Sensor-SDK/blob/61951daac782234f4f28322c0904ba1c4702d0ba/src/transformation/rgbz.c#L738-L743

and

https://github.com/microsoft/Azure-Kinect-Sensor-SDK/blob/61951daac782234f4f28322c0904ba1c4702d0ba/src/transformation/rgbz.c#L843

Updated code could be similar to...

k4a_image_format_t custom_format = K4A_IMAGE_FORMAT_CUSTOM8;
if ((custom_image_descriptor->format == K4A_IMAGE_FORMAT_CUSTOM16) ||
    (custom_image_descriptor->format == K4A_IMAGE_FORMAT_IR16))
{
    custom_bytes_per_pixel = 2;
    custom_format = custom_image_descriptor->format;
}

...and...

else if ((custom_image_descriptor->format == K4A_IMAGE_FORMAT_CUSTOM16) ||
         (custom_image_descriptor->format == K4A_IMAGE_FORMAT_IR16))

Why Fix?

• Reduces issues to manage like this one you are reading and #1188.
• Reduces customer dev/test churn to use transform API in a way they expect it to work, but the API doesn't work
• Reduces CPU and memory usage. The workaround to k4a::image::create_from_buffer() wrap around the already existing IR16 is wasteful of both CPU and memory.
• Doesn't break any compatibility.