You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
By downloading the OSI data you provided, I found that DSM data is in uint16 format, so do you still need the following processing in dataloader? depth = cv2.imread(depth_name,-1); depth = (depth*1000).astype(np.uint16)
I understand that depth = self.to_tensor(depth)/100000 in Totensor(), which you define in Transform, is to regularize all the values in the DSM and set them between 0 and 1. I do not know whether I understand correctly. If I am wrong, I hope to get your explanation.
However, the above processing is actually carried out under the premise of default depth*1000<65535, but I see that the DSM of OSI dataset provided by you is all in uint16 format, and the maximum value is 65531. For this situation, how do you design the dataloader of OSI data set.
Hope to get your answer.:)
The text was updated successfully, but these errors were encountered:
By downloading the OSI data you provided, I found that DSM data is in uint16 format, so do you still need the following processing in dataloader?
depth = cv2.imread(depth_name,-1); depth = (depth*1000).astype(np.uint16)
I understand that
depth = self.to_tensor(depth)/100000
in Totensor(), which you define in Transform, is to regularize all the values in the DSM and set them between 0 and 1. I do not know whether I understand correctly. If I am wrong, I hope to get your explanation.However, the above processing is actually carried out under the premise of default depth*1000<65535, but I see that the DSM of OSI dataset provided by you is all in uint16 format, and the maximum value is 65531. For this situation, how do you design the dataloader of OSI data set.
Hope to get your answer.:)
The text was updated successfully, but these errors were encountered: