Improve simgetimages speed

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

@@ -28,9 +28,8 @@
 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_rgba = image1d.reshape(image.height, image.width, 4)
-    image_rgba = np.flipud(image_rgba)
-    return image_rgba[:, :, 0:3]
+    image_rgb = image1d.reshape(image.height, image.width, 3)
+    return image_rgb
 
 while (True):
     car_state = client.getCarState()
@@ -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()
@@ -53,7 +53,7 @@
         airsim.ImageRequest("0", airsim.ImageType.DepthVis),  #depth visualization image
         airsim.ImageRequest("1", airsim.ImageType.DepthPerspective, True), #depth in perspective projection
         airsim.ImageRequest("1", airsim.ImageType.Scene), #scene vision image in png format
-        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGBA array
+        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGB array
     print('Retrieved images: %d', len(responses))
 
     for response in responses:
@@ -68,12 +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_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 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

@@ -40,7 +40,7 @@
     ImageRequest(0, airsim.AirSimImageType.DepthVis),  #depth visualiztion image
     ImageRequest(1, airsim.AirSimImageType.DepthPerspective, True), #depth in perspective projection
     ImageRequest(1, airsim.AirSimImageType.Scene), #scene vision image in png format
-    ImageRequest(1, airsim.AirSimImageType.Scene, False, False)])  #scene vision image in uncompressed RGBA array
+    ImageRequest(1, airsim.AirSimImageType.Scene, False, False)])  #scene vision image in uncompressed RGB array
 print('Retrieved images: %d' % len(responses))
 
 tmp_dir = os.path.join(tempfile.gettempdir(), "airsim_drone")
@@ -64,10 +64,8 @@
     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 fliped vertically
-        img_rgba[:,:,1:2] = 100 #just for fun add little bit of green in all pixels
-        AirSimClientBase.write_png(os.path.normpath(filename + '.greener.png'), img_rgba) #write to png
+        img_rgb = img1d.reshape(response.height, response.width, 3) #reshape array to 3 channel image array H X W X 3
+        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
 """
@@ -91,11 +91,11 @@
     # get camera images from the car
     responses1 = client.simGetImages([
         airsim.ImageRequest("0", airsim.ImageType.DepthVis),  #depth visualization image
-        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], "Car1")  #scene vision image in uncompressed RGBA array
+        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], "Car1")  #scene vision image in uncompressed RGB array
     print('Car1: Retrieved images: %d' % (len(responses1)))
     responses2 = client.simGetImages([
         airsim.ImageRequest("0", airsim.ImageType.Segmentation),  #depth visualization image
-        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], "Car2")  #scene vision image in uncompressed RGBA array
+        airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], "Car2")  #scene vision image in uncompressed RGB array
     print('Car2: Retrieved images: %d' % (len(responses2)))
 
     for response in responses1 + responses2:
@@ -109,12 +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_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 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()
@@ -54,20 +54,10 @@
     else: #uncompressed array - numpy demo
         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
-        #airsim.write_png(os.path.normpath(filename + '.numpy.png'), img_rgba) #write to png 
+        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
 
         #find unique colors
-        print(np.unique(img_rgba[:,:,0], return_counts=True)) #red
-        print(np.unique(img_rgba[:,:,1], return_counts=True)) #green
-        print(np.unique(img_rgba[:,:,2], return_counts=True)) #blue  
-        print(np.unique(img_rgba[:,:,3], return_counts=True)) #blue
-
-
-
-
-
-
-
-
+        print(np.unique(img_rgb[:,:,0], return_counts=True)) #red
+        print(np.unique(img_rgb[:,:,1], return_counts=True)) #green
+        print(np.unique(img_rgb[:,:,2], return_counts=True)) #blue  

PythonClient/imitation_learning/drive_model.py

@@ -39,8 +39,8 @@ def get_image():
     """
     image_response = client.simGetImages([airsim.ImageRequest("0", airsim.ImageType.Scene, False, False)])[0]
     image1d = np.fromstring(image_response.image_data_uint8, dtype=np.uint8)
-    image_rgba = image1d.reshape(image_response.height, image_response.width, 4)
-    return image_rgba[78:144,27:227,0:3].astype(float)
+    image_rgb = image1d.reshape(image_response.height, image_response.width, 3)
+    return image_rgb[78:144,27:227,0:2].astype(float)
 
 while True:    
     # Update throttle value according to steering angle

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/kinect_publisher.py

@@ -40,8 +40,8 @@ def getDepthImage(self,response_d):
 
     def getRGBImage(self,response_rgb):
         img1d = np.fromstring(response_rgb.image_data_uint8, dtype=np.uint8)
-        img_rgba = img1d.reshape(response_rgb.height, response_rgb.width, 4)
-        img_rgb = img_rgba[..., :3][..., ::-1]
+        img_rgb = img1d.reshape(response_rgb.height, response_rgb.width, 3)
+        img_rgb = img_rgb[..., :3][..., ::-1]
         return img_rgb
 
     def enhanceRGB(self,img_rgb):

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
 """
@@ -59,11 +59,11 @@
 # get camera images from the car
 responses1 = client.simGetImages([
     airsim.ImageRequest("0", airsim.ImageType.DepthVis),  #depth visualization image
-    airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], vehicle_name="Drone1")  #scene vision image in uncompressed RGBA array
+    airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], vehicle_name="Drone1")  #scene vision image in uncompressed RGB array
 print('Drone1: Retrieved images: %d' % len(responses1))
 responses2 = client.simGetImages([
     airsim.ImageRequest("0", airsim.ImageType.DepthVis),  #depth visualization image
-    airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], vehicle_name="Drone2")  #scene vision image in uncompressed RGBA array
+    airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)], vehicle_name="Drone2")  #scene vision image in uncompressed RGB array
 print('Drone2: Retrieved images: %d' % len(responses2))
 
 tmp_dir = os.path.join(tempfile.gettempdir(), "airsim_drone")
@@ -87,10 +87,8 @@
     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
+        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
 
 airsim.wait_key('Press any key to reset to original state')
 

PythonClient/ros/car_image_raw.py

@@ -22,21 +22,21 @@ def airpub():
     while not rospy.is_shutdown():
          # get camera images from the car
         responses = client.simGetImages([
-            airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGBA array
+            airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGB array
 
         for response in responses:
-            img_rgba_string = response.image_data_uint8
+            img_rgb_string = response.image_data_uint8
 
         # Populate image message
         msg=Image() 
         msg.header.stamp = rospy.Time.now()
         msg.header.frame_id = "frameId"
-        msg.encoding = "rgba8"
+        msg.encoding = "rgb8"
         msg.height = 360  # resolution should match values in settings.json 
         msg.width = 640
-        msg.data = img_rgba_string
+        msg.data = img_rgb_string
         msg.is_bigendian = 0
-        msg.step = msg.width * 4
+        msg.step = msg.width * 3
 
         # log time and size of published image
         rospy.loginfo(len(response.image_data_uint8))

PythonClient/ros/drone_image_raw.py

@@ -23,21 +23,21 @@ def airpub():
     while not rospy.is_shutdown():
          # get camera images from the car
         responses = client.simGetImages([
-            airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGBA array
+            airsim.ImageRequest("1", airsim.ImageType.Scene, False, False)])  #scene vision image in uncompressed RGB array
 
         for response in responses:
-            img_rgba_string = response.image_data_uint8
+            img_rgb_string = response.image_data_uint8
 
         # Populate image message
         msg=Image() 
         msg.header.stamp = rospy.Time.now()
         msg.header.frame_id = "frameId"
-        msg.encoding = "rgba8"
+        msg.encoding = "rgb8"
         msg.height = 360  # resolution should match values in settings.json 
         msg.width = 640
-        msg.data = img_rgba_string
+        msg.data = img_rgb_string
         msg.is_bigendian = 0
-        msg.step = msg.width * 4
+        msg.step = msg.width * 3
 
         # log time and size of published image
         rospy.loginfo(len(response.image_data_uint8))