Show Posts
1
Python and Java API / Adding markers for georeferencing of an orthomosaic
« on: September 17, 2020, 05:30:04 PM »
I have looked through a number of threads and pieced together what I thought should work for importing markers and assigning them their image and real world coordinates. However, after running the script I fear that something is wrong since when I check the reference tab in the GUI, it doesn't show that my markers have their real world coordinates assigned.
Here is the code I am using:
One additional question to this. If I add the markers like this, can I then later use the reference_preselection mode when I am trying to match the images? If so, would the following be correct (after having added the markers as above):
Thanks!
Here is the code I am using:
Code: [Select]
import Metashape as MS
# set the coordinate ref system of the markers
crs_ = MS.CoordinateSystem("EPSG::%d" % crs)
chunk.marker_crs = crs_
# set the z-accuracy very high because we don't have a good measure for this
chunk.marker_location_accuracy = [0.05, 0.05, 1000]
file = open('./export.csv', "rt")#input file
# skips header
lines = file.readlines()[1:]
for line in lines:
line = line.strip('\n')
sp_line = line.rsplit(separation, 6) #splitting line
# Strip extra quotation marks if they exist
for i, item in enumerate(sp_line):
sp_line[i] = item.strip('"')
camera_name = sp_line[0]
marker_name = sp_line[1]
x = float(sp_line[2]) #x-coord in pixels
y = float(sp_line[3]) #y-coord in pixels
cx = float(sp_line[4]) #world x-coord of marker
cy = float(sp_line[5]) #world y-coord of marker
cz = float(sp_line[6]) #world z-coord of marker
# flag to indicate if marker is found
flag = 0
# Create a Projection from the pixel coordinates
pixel_coord = MS.Marker.Projection(MS.Vector([x, y]), True)
# get the camera with same name as the name in current line
for cam in chunk.cameras:
if cam.label == camera_name:
#searching for a marker with the same ID as in txt file
for marker in chunk.markers:
if marker.label == marker_name:
# assign the marker's x,y coords in the current cam
marker.projections[cam] = pixel_coord
flag = 1
break
# if marker ID not yet existing in the chunk then create
if flag == 0:
marker = chunk.addMarker()
marker.label = marker_name
# assign the marker's x,y coords in the current cam
marker.projections[cam] = pixel_coord
# Add the world coordinates to the marker
marker.Reference.location = MS.Vector([cx, cy, cz])
# Set the accuracy of the marker's coords
marker.Reference.accuracy = MS.Vector(accuracy)
# enable the marker
marker.Reference.enabled = True
# move onto the next line of the text file
break
file.close()
print("Import finished.")
One additional question to this. If I add the markers like this, can I then later use the reference_preselection mode when I am trying to match the images? If so, would the following be correct (after having added the markers as above):
Code: [Select]
chunk.crs = chunk.marker_crs
chunk.updateTransform()
chunk.matchPhotos(generic_preselection=False, reference_preselection=True)
Thanks!
