How to process realsense 's IR image

[zhiguo-rbs]

I read realsense IR image which is gray image, not RGB image, how to process it to forward the model?

[steven-mantis]

Yesterday I tried to convert the gray image to rgb888 image with opencv cv::cvtColor and feed it to the model and finally get the depthmap.

[zhiguo-rbs]

I tried to repeat the gray image along channel axis, it is also work
It does the same with you @steven-mantis

[abdelrahman-gaber]

Repeating the grayscale frame 3 times should work. I used the code below and it is working fine:

img0 = cv2.cvtColor(infrared_image_0, cv2.COLOR_GRAY2RGB)
img1 = cv2.cvtColor(infrared_image_1, cv2.COLOR_GRAY2RGB)

You need to enable the infrared streams properly as mentioned here then convert them to numpy arrays as follows:

infrared_frame_0 = frames.get_infrared_frame(1)
infrared_frame_1 = frames.get_infrared_frame(2)

infrared_image_0 = np.asanyarray(infrared_frame_0.get_data())
infrared_image_1 = np.asanyarray(infrared_frame_1.get_data())