Changing depth resolution different from 256x144

[DanieleMarchisotti]

Bug report

• AirSim Version/#commit: Linux 1.8.1
• UE/Unity version: 4.27
• autopilot version: N/A
• OS Version: Ubuntu 22.04

What's the issue you encountered?

Slowing dramatically the number of depth images saved if the resolution is different from 256x144.
I can save 1 image per second with 256x144 of resolution, but when the resolution is changed I can save only 1 image every 30 s.

Settings

This is the portion of code I use to save images. I pause the simulation to get synchronized depth and rgb images.

...
    # Request scene and segmentation images separately
    scene_requests = [
        airsim.ImageRequest("0", airsim.ImageType.Scene, False, False), # camera_name, image_type, pixels_as_float, compress
        airsim.ImageRequest("0", airsim.ImageType.DepthPlanar, True, False), # camera_name, image_type, pixels_as_float, compress
    ]
    frame_num = 0
    while capturing:
        try:
            # pausing simulation to capture data synchronously
            client_sim.call('simPause', True)
            scene_responses = client.simGetImages(scene_requests)
            for idx, response in enumerate(scene_responses):
                if response.pixels_as_float:
                    filename = save_dir + "/depth/" + str(response.time_stamp)
                    print("Type %d, size %d" % (response.image_type, len(response.image_data_float)))
                    airsim.write_pfm(os.path.normpath(filename + '.pfm'), airsim.get_pfm_array(response))
                else: #uncompressed array
                    filename = save_dir + "/rgb/" + str(response.time_stamp)
                    print("Type %d, size %d" % (response.image_type, len(response.image_data_uint8)))
                    img1d = np.fromstring(response.image_data_uint8, dtype=np.uint8) # get numpy array
                    img_rgb = img1d.reshape(response.height, response.width, 3)
                    img_rgb = cv2.cvtColor(img_rgb, cv2.COLOR_BGR2RGB)
                    cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png
            # resuming simulation after capture
            client_sim.call('simPause', False)
            print(f"Captured frame: ", str(frame_num))
            frame_num += 1
            time.sleep(1)
        except Exception as e:
            print(f"Failed to capture image: {str(e)}")
...

settings.json:
If I change the resolution I get the issue.

{
    "SettingsVersion": 1.2,
    "SimMode": "Multirotor",
    "Vehicles": {
      "drone_test": {
        "VehicleType": "SimpleFlight",
        "AutoCreate": true,
        "X": 0,
        "Y": 0,
        "Z": 0,
        "Cameras": {
          "0": {
            "X": 0.5,
            "Y": 0,
            "Z": 0.1,
            "Pitch": -30.0,
            "Roll": 0.0,
            "Yaw": 0.0,
            "CaptureSettings": [
              { "ImageType": 0,
                "Width": 1280,
                "Height": 720,
                "FOV_Degrees": 32
              },
              { "ImageType": 1,
              "Width": 256,
              "Height": 144,
              "FOV_Degrees": 32
              },
              { "ImageType": 7,
              "Width": 1280,
              "Height": 720,
              "FOV_Degrees": 32
              }
            ]
          }
        }
      }
    },
    "SubWindows": [
      {
        "WindowID": 0,
        "CameraName": "0",
        "ImageType": 0,
        "VehicleName": "drone_test",
        "Visible": true
      },
      {
        "WindowID": 1,
        "CameraName": "0",
        "ImageType": 1,
        "VehicleName": "drone_test",
        "Visible": true
      },
      {
        "WindowID": 2,
        "CameraName": "0",
        "ImageType": 7,
        "VehicleName": "drone_test",
        "Visible": true
      }
    ]
  }

How can the issue be reproduced?

Run the code I put above with the settigns.json above

Include full error message in text form

No error message but just slowing down the saving of depth images.
I can save only one image in approximately 30 seconds.