Why picture pixels cannot be converted to floating-point numbers

[LLalang]

responses = client.simGetImages([airsim.ImageRequest("0", airsim.ImageType.DepthPerspective, True, True)])
current_state = transform_input(responses)

def transform_input(responses):

img1d = np.array(responses[0].image_data_float, dtype=np.float)

img1d = 255/np.maximum(np.ones(img1d.size), img1d)



img2d = np.reshape(img1d, (responses[0].height, responses[0].width))

from PIL import Image
image = Image.fromarray(img2d)
im_final = np.array(image.resize((84, 84)).convert('L')) 

return im_final

[LLalang]

Has anyone ever encountered such a problem？

[rallen10]

I have the same problem. image_data_float is [0.0] even though image_data_uint8 has large amounts of data. I am using pixels_as_float=True. Did you ever find a solution?

[stale]

This issue has been automatically marked as stale because it has not had activity from the community in the last year. It will be closed if no further activity occurs within 20 days.

[sandilyasg]

LLalang were you able to resolve this issue? I am using Unity and image_data_float always returns an empty image. However, when I view the depth, semantic segmentation, and RGB images directly in Unity, they appear to be working fine.