使用GRAYSCALE图像创建BGR图像

Question

Camera provides GRAYSCALE data while deep learning model requires BGR. Is it possible to create artificial bgr image by "stacking" the grayscale image using opencv or numpy or anything?

I am using two ROS packages created by other people. I have tried the following:

Attempted altering the model to accept grayscale as described by the original repo.

Attempted to convert grayscale to bgr using

np_image = cv2.cvtColor(np_image, cv2.COLOR_GRAY2BGR)

Attempted to create an array consisting of the image times three

np_image = np.array([np_image, np_image, np_image])

The following code retrieves the image

    def run(self):
        self._result_pub = rospy.Publisher('~result', Result, queue_size=1)
        vis_pub = rospy.Publisher('~visualization', Image, queue_size=1)
        sub = rospy.Subscriber('~input', Image,
                               self._image_callback, queue_size=1) 

        rate = rospy.Rate(self._publish_rate)
        while not rospy.is_shutdown():
            if self._msg_lock.acquire(False):
                msg = self._last_msg
                self._last_msg = None
                self._msg_lock.release()
            else:
                rate.sleep()
                continue

            if msg is not None:
                np_image = self._cv_bridge.imgmsg_to_cv2(msg, 
                    desired_encoding='mono8')
                #np_image = cv2.cvtColor(np_image, cv2.COLOR_GRAY2BGR) #(fail)
                np_image = np.array([np_image, np_image, np_image])
                # Run detection
                results = self._model.detect([np_image], verbose=0)
                result = results[0]
                result_msg = self._build_result_msg(msg, result)
                self._result_pub.publish(result_msg)

def _image_callback(self, msg):
        rospy.logdebug("Get an image")
        if self._msg_lock.acquire(False):
            self._last_msg = msg
            self._msg_lock.release()

I expect the model to retrieve the data, evaluate it and return an object id and mask. However, I get the following error:

Traceback (most recent call last):
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/nodes/mask_rcnn_node", line 196, in <module>
    main()
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/nodes/mask_rcnn_node", line 192, in main
    node.run()
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/nodes/mask_rcnn_node", line 116, in run
    results = self._model.detect([np_image], verbose=0)
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/src/mask_rcnn_ros/model.py", line 2333, in detect
    molded_images, image_metas, windows = self.mold_inputs(images)
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/src/mask_rcnn_ros/model.py", line 2236, in mold_inputs
    padding=self.config.IMAGE_PADDING)
  File "/home/riwo-rack-pc/ROS_Mask_rcnn/src/mask_rcnn_ros/src/mask_rcnn_ros/utils.py", line 409, in resize_image
    image, (round(h * scale), round(w * scale)))
  File "/home/riwo-rack-pc/.local/lib/python2.7/site-packages/scipy/misc/pilutil.py", line 490, in imresize
    imnew = im.resize(size, resample=func[interp])
  File "/home/riwo-rack-pc/.local/lib/python2.7/site-packages/PIL/Image.py", line 1806, in resize
    return self._new(self.im.resize(size, resample, box))
TypeError: integer argument expected, got float

I followed the trace as best I could but can't find the point where the float is introduced.

Answer 1

merged_image = cv2.merge((np_image, np_image, np_image))

应该给你你想要的东西。