Topics

1

Python and Java API / Parallelizing large scale computations

« on: December 08, 2022, 10:54:45 AM »

Hi everyone, I have an application that I am trying to make more computationally efficient.

I would like to project ALL dense cloud points to ALL cameras in my project to see which 3D points have valid projections.  So, I loop over all cameras and run a camera.project on all the 3D points.

The loop over all cameras is a simple for loop, but I have been able to speed up the loop over all 3D points by using a map operation over camera.project.

For each camera, on my M1 Max Mac with 64 GB RAM and 10 cores, the projection of all 3D points for each camera takes about 5 seconds.  A typical project has ~10 million points and ~300 cameras, so this operation takes a few minutes, which is something I would like to improve.

Some of the things I do now:
1. Remove redundant 3D points written out in the dense cloud, cuts down up to a third of 3D points (I don't know why the redundant points get written in the first place).
2. Instead of looping over every 3D point, I use a map operation to execute camera.project over all the 3D points.

Question:
1. How can I parallelize the camera.project step over all 3D points?  I am assuming that it is currently running only on 1 core
2. How can I parallelize the loop over all the cameras?

Thanks in advance, any help is greatly appreciated.

2

Python and Java API / cam.project, cam.calibration.project and cam.sensor.calibration.project

« on: November 30, 2022, 10:25:39 AM »

Are these three commands supposed to produce the same output, i.e. the 2D projection of a 3D point on an image?

I am getting the same output for cam.calibration.project and cam.sensor.calibration.project, but cam.sensor returns empty for the same exact point.

In [25]: cam.project(temp_b[0])

In [26]: cam.calibration.project(temp_b[0])
Out[26]: 2022-11-30 12:53:57 Vector([226.91737982171833, 3066.876461424863])

In [27]: cam.sensor.calibration.project(temp_b[0])
Out[27]: 2022-11-30 12:54:04 Vector([226.91737982171833, 3066.876461424863])

Thanks in advance for your help.

3

Python and Java API / Marker pixel location from a single image without alignment

« on: July 21, 2022, 02:42:33 PM »

I have a single image (actually a Metashape Orthophoto) which has some markers that I can run marker detection on.  All I want are the pixel coordinates of these markers, but I cannot use the markers.projections approach because there is no alignment to speak of here and projections returns None.

I tried to export the markers to an XML file (below) and sure enough, I can see the locations of two markers in pixel coordinates towards the end of the file.  How do I recover these using the Python API? 

Thank you.

---------------------
<?xml version="1.0" encoding="UTF-8"?>
<document version="1.5.0">
  <chunk label="Chunk 1" enabled="true">
    <sensors next_id="1">
      <sensor id="0" label="unknown" type="frame">
        <resolution width="3038" height="3561"/>
        <property name="layer_index" value="0"/>
        <bands>
          <band label="Red"/>
          <band label="Green"/>
          <band label="Blue"/>
        </bands>
        <data_type>uint16</data_type>
      </sensor>
    </sensors>
    <components next_id="0"/>
    <cameras next_id="1" next_group_id="0">
      <camera id="0" sensor_id="0" label="Austin_6_24_2_test_pib-bchudzsz"/>
    </cameras>
    <markers next_id="2" next_group_id="0">
      <marker id="0" label="target 116"/>
      <marker id="1" label="target 117"/>
    </markers>
    <reference>LOCAL_CS["Local Coordinates (m)",LOCAL_DATUM["Local Datum",0],UNIT["metre",1,AUTHORITY["EPSG","9001"]]]</reference>
    <settings>
      <property name="accuracy_tiepoints" value="1"/>
      <property name="accuracy_cameras" value="10"/>
      <property name="accuracy_cameras_ypr" value="10"/>
      <property name="accuracy_markers" value="0.0050000000000000001"/>
      <property name="accuracy_scalebars" value="0.001"/>
      <property name="accuracy_projections" value="0.5"/>
    </settings>
    <frames next_id="1">
      <frame id="0">
        <markers>
          <marker marker_id="0">
            <location camera_id="0" pinned="true" x="2619.54663" y="1992.58142"/>
          </marker>
          <marker marker_id="1">
            <location camera_id="0" pinned="true" x="2820.41455" y="2001.7561"/>
          </marker>
        </markers>
      </frame>
    </frames>
  </chunk>
</document>