[simgetimages] fix bgr<->rgb in renderrequest

remove old code's np.flipuds
replace airsim.write_png with cv2.imwrite

PythonClient/airsim/pfm.py

@@ -49,7 +49,6 @@ def read_pfm(file):
 
     data = np.reshape(data, shape)
     # DEY: I don't know why this was there.
-    #data = np.flipud(data)
     file.close()
 
     return data, scale
@@ -64,8 +63,6 @@ def write_pfm(file, image, scale=1):
     if image.dtype.name != 'float32':
         raise Exception('Image dtype must be float32.')
 
-    image = np.flipud(image)
-
     if len(image.shape) == 3 and image.shape[2] == 3: # color image
         color = True
     elif len(image.shape) == 2 or len(image.shape) == 3 and image.shape[2] == 1: # greyscale

PythonClient/airsim/utils.py

@@ -157,7 +157,6 @@ def read_pfm(file):
 
     data = np.reshape(data, shape)
     # DEY: I don't know why this was there.
-    #data = np.flipud(data)
     file.close()
 
     return data, scale
@@ -172,11 +171,9 @@ def write_pfm(file, image, scale=1):
     if image.dtype.name != 'float32':
         raise Exception('Image dtype must be float32.')
 
-    image = np.flipud(image)
-
     if len(image.shape) == 3 and image.shape[2] == 3: # color image
         color = True
-    elif len(image.shape) == 2 or len(image.shape) == 3 and image.shape[2] == 1: # greyscale
+    elif len(image.shape) == 2 or len(image.shape) == 3 and image.shape[2] == 1: # grayscale
         color = False
     else:
         raise Exception('Image must have H x W x 3, H x W x 1 or H x W dimensions.')
@@ -206,9 +203,9 @@ def write_png(filename, image):
     height = image.shape[0]
 
     # reverse the vertical line order and add null bytes at the start
-    width_byte_4 = width * 4
-    raw_data = b''.join(b'\x00' + buf[span:span + width_byte_4]
-                        for span in range((height - 1) * width_byte_4, -1, - width_byte_4))
+    width_byte_3 = width * 3
+    raw_data = b''.join(b'\x00' + buf[span:span + width_byte_3]
+                        for span in range((height - 1) * width_byte_3, -1, - width_byte_3))
 
     def png_pack(png_tag, data):
         chunk_head = png_tag + data

PythonClient/car/drive_straight.py

@@ -29,7 +29,6 @@ def get_image():
     image = client.simGetImages([airsim.ImageRequest("0", airsim.ImageType.Scene, False, False)])[0]
     image1d = np.fromstring(image.image_data_uint8, dtype=np.uint8)
     image_rgb = image1d.reshape(image.height, image.width, 3)
-    image_rgb = np.flipud(image_rgb)
     return image_rgb
 
 while (True):
@@ -42,13 +41,11 @@ def get_image():
     else:
         car_controls.throttle = 0.0
 
-    #image_buf[0] = get_image()
     #state_buf[0] = np.array([car_controls.steering, car_controls.throttle, car_controls.brake, car_state.speed])
     #model_output = model.predict([image_buf, state_buf])
     #car_controls.steering = float(model_output[0][0])
     car_controls.steering = 0
 
     print('Sending steering = {0}, throttle = {1}'.format(car_controls.steering, car_controls.throttle))
 
-    client.setCarControls(car_controls)
-
+    client.setCarControls(car_controls)

PythonClient/car/hello_car.py

@@ -1,9 +1,9 @@
-import setup_path 
 import airsim
-
-import time
-import os
+import cv2
 import numpy as np
+import os
+import setup_path 
+import time
 
 # connect to the AirSim simulator 
 client = airsim.CarClient()
@@ -68,11 +68,9 @@
             airsim.write_file(os.path.normpath(filename + '.png'), response.image_data_uint8)
         else: #uncompressed array
             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) #reshape array to 3 channel image array H X W X 3
-            img_rgb = np.flipud(img_rgb) #original image is flipped vertically
-            airsim.write_png(os.path.normpath(filename + '.png'), img_rgb) #write to png 
-
+            img1d = np.fromstring(response.image_data_uint8, dtype=np.uint8) # get numpy array
+            img_rgb = img1d.reshape(response.height, response.width, 3) # reshape array to 3 channel image array H X W X 3
+            cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png 
 
 #restore to original state
 client.reset()

PythonClient/car/legacy_hello_car.py

@@ -65,7 +65,6 @@
         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) #reshape array to 3 channel image array H X W X 3
-        img_rgb = np.flipud(img_rgb) #original image is fliped vertically
         AirSimClientBase.write_png(os.path.normpath(filename + '.png'), img_rgb) #write to png
 
 AirSimClientBase.wait_key('Press any key to reset to original state')

PythonClient/car/multi_agent_car.py

@@ -1,9 +1,9 @@
-import setup_path 
 import airsim
-
-import time
-import os
+import cv2
 import numpy as np
+import os
+import setup_path 
+import time
 
 # Use below in settings.json with blocks environment
 """
@@ -109,11 +109,9 @@
             airsim.write_file(os.path.normpath(filename + '.png'), response.image_data_uint8)
         else: #uncompressed array
             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) #reshape array to 3 channel image array H X W X 3
-            img_rgb = np.flipud(img_rgb) #original image is flipped vertically
-            airsim.write_png(os.path.normpath(filename + '.png'), img_rgb) #write to png 
-
+            img1d = np.fromstring(response.image_data_uint8, dtype=np.uint8) # get numpy array
+            img_rgb = img1d.reshape(response.height, response.width, 3) # reshape array to 3 channel image array H X W X 3
+            cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png
 
 #restore to original state
 client.reset()

PythonClient/computer_vision/segmentation.py

@@ -1,10 +1,10 @@
 # In settings.json first activate computer vision mode: 
 # https://github.com/Microsoft/AirSim/blob/master/docs/image_apis.md#computer-vision-mode
 
-import setup_path 
 import airsim
-
+import cv2
 import numpy as np
+import setup_path 
 
 client = airsim.VehicleClient()
 client.confirmConnection()
@@ -55,8 +55,7 @@
         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) #reshape array to 3 channel image array H X W X 3
-        img_rgb = np.flipud(img_rgb) #original image is flipped vertically
-        #airsim.write_png(os.path.normpath(filename + '.numpy.png'), img_rgb) #write to png 
+        # cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png
 
         #find unique colors
         print(np.unique(img_rgb[:,:,0], return_counts=True)) #red

PythonClient/multirotor/hello_drone.py

@@ -5,6 +5,7 @@
 import os
 import tempfile
 import pprint
+import cv2
 
 # connect to the AirSim simulator
 client = airsim.MultirotorClient()
@@ -59,11 +60,9 @@
         airsim.write_file(os.path.normpath(filename + '.png'), response.image_data_uint8)
     else: #uncompressed array
         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_rgba = img1d.reshape(response.height, response.width, 4) #reshape array to 4 channel image array H X W X 4
-        img_rgba = np.flipud(img_rgba) #original image is flipped vertically
-        img_rgba[:,:,1:2] = 100 #just for fun add little bit of green in all pixels
-        airsim.write_png(os.path.normpath(filename + '.greener.png'), img_rgba) #write to png
+        img1d = np.fromstring(response.image_data_uint8, dtype=np.uint8) # get numpy array
+        img_rgb = img1d.reshape(response.height, response.width, 3) # reshape array to 4 channel image array H X W X 3
+        cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png
 
 airsim.wait_key('Press any key to reset to original state')
 

PythonClient/multirotor/multi_agent_drone.py

@@ -1,10 +1,10 @@
-import setup_path 
 import airsim
-
+import cv2
 import numpy as np
 import os
-import tempfile
 import pprint
+import setup_path 
+import tempfile
 
 # Use below in settings.json with Blocks environment
 """
@@ -88,8 +88,7 @@
         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) #reshape array to 3 channel image array H X W X 3
-        img_rgb = np.flipud(img_rgb) #original image is flipped vertically
-        airsim.write_png(os.path.normpath(filename + '.png'), img_rgb) #write to png
+        cv2.imwrite(os.path.normpath(filename + '.png'), img_rgb) # write to png
 
 airsim.wait_key('Press any key to reset to original state')
 

Unreal/Plugins/AirSim/Source/RenderRequest.cpp

@@ -112,9 +112,9 @@ void RenderRequest::getScreenshot(std::shared_ptr<RenderParams> params[], std::v
                 else {
                     uint8* ptr = results[i]->image_data_uint8.GetData();
                     for (const auto& item : results[i]->bmp) {
-                        *ptr++ = item.R;
-                        *ptr++ = item.G;
                         *ptr++ = item.B;
+                        *ptr++ = item.G;
+                        *ptr++ = item.R;
                     }
                 }
             }