pose2sim/Pose2Sim/Utilities/calib_qca_to_toml.py

238 lines
8.0 KiB
Python
Raw Normal View History

2023-07-19 17:37:20 +08:00
#!/usr/bin/env python
# -*- coding: utf-8 -*-
'''
##################################################
## QCA CALIBRATION TO TOML CALIBRATION ##
##################################################
Convert a Qualisys .qca.txt calibration file
to an OpenCV .toml calibration file
Usage:
from Pose2Sim.Utilities import calib_qca_to_toml; calib_qca_to_toml.calib_qca_to_toml_func(r'<input_qca_file>')
2023-09-21 23:39:28 +08:00
OR python -m calib_qca_to_toml -i input_qca_file
OR python -m calib_qca_to_toml -i input_qca_file --binning_factor 2 -o output_toml_file
2023-07-19 17:37:20 +08:00
'''
## INIT
import os
import argparse
import re
import numpy as np
from lxml import etree
import cv2
## AUTHORSHIP INFORMATION
__author__ = "David Pagnon"
__copyright__ = "Copyright 2021, Pose2Sim"
__credits__ = ["David Pagnon"]
__license__ = "BSD 3-Clause License"
2024-07-10 16:12:57 +08:00
__version__ = "0.9.4"
2023-07-19 17:37:20 +08:00
__maintainer__ = "David Pagnon"
__email__ = "contact@david-pagnon.com"
__status__ = "Development"
## FUNCTIONS
Pose estimation test (#116) Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work! General: - Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend - The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version. - Added logging information to pose estimation - Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust) - Readme.md: added information about PoseEstimation with RTMLib - added poseEstimation to tests.py - created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well Json files: - the json files only saved one person, I made it save all the detected ones - tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge - took the save_to_openpose function out from the main function - minified the json files (they take less space when all spaces are removed) Detection results: - Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it. Others in Config.toml: - removed the "to_openpose" option, which is not needed - added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images'] - changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation) - added the flag: time_range_around_maxspeed in synchronization - automatically detect framerate from video, or set to 60 fps if we work from images (or give a value) - frame_range -> time_range - moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling) - removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead Others: - Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip) - scaling from motion instead of static pose (will have to study whether it's as good or not) - added logging to synchronization - Struggled quite a bit with continuous integration * Starting point of integrating RTMPose into Pose2Sim. (#111) * RTM_to_Open Convert format from RTMPose to OpenPose * rtm_intergrated * rtm_integrated * rtm_integrated * rtm_integrated * rtm * Delete build/lib/Pose2Sim directory * rtm * Delete build/lib/Pose2Sim directory * Delete onnxruntime-gpu * device = cpu * add pose folder * Update tests.py * added annotation * fix typo * Should work be still lots of tests to run. Detailed commit coming soon * intermediary commit * last checks before v0.9.0 * Update continuous-integration.yml * Update tests.py * replaced tabs with spaces * unittest issue * unittest typo * deactivated display for CI test of pose detection * Try to make continuous integration work * a * b * c * d * e * f * g * h * i * j * k * l --------- Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
def natural_sort_key(s):
"""
Key for natural sorting of strings containing numbers.
"""
return [int(c) if c.isdigit() else c.lower() for c in re.split(r'(\d+)', s)]
2023-07-19 17:37:20 +08:00
def read_qca(qca_path, binning_factor):
'''
Read a Qualisys .qca.txt calibration file
Returns 5 lists of size N (N=number of cameras):
- ret: residual reprojection error in _mm_: list of floats
- C (camera name),
- S (image size),
- D (distorsion),
- K (intrinsic parameters),
- R (extrinsic rotation),
- T (extrinsic translation)
'''
root = etree.parse(qca_path).getroot()
ret, C, S, D, K, R, T = [], [], [], [], [], [], []
vid_id = []
# Camera name
for i, tag in enumerate(root.findall('cameras/camera')):
ret += [float(tag.attrib.get('avg-residual'))/1000]
C += [tag.attrib.get('serial')]
if tag.attrib.get('model') in ('Miqus Video', 'Miqus Video UnderWater', 'none'):
vid_id += [i]
# Image size
for tag in root.findall('cameras/camera/fov_video'):
w = (float(tag.attrib.get('right')) - float(tag.attrib.get('left'))) /binning_factor
h = (float(tag.attrib.get('bottom')) - float(tag.attrib.get('top'))) /binning_factor
S += [[w, h]]
# Intrinsic parameters: distorsion and intrinsic matrix
for i, tag in enumerate(root.findall('cameras/camera/intrinsic')):
k1 = float(tag.get('radialDistortion1'))/64/binning_factor
k2 = float(tag.get('radialDistortion2'))/64/binning_factor
p1 = float(tag.get('tangentalDistortion1'))/64/binning_factor
p2 = float(tag.get('tangentalDistortion2'))/64/binning_factor
D+= [np.array([k1, k2, p1, p2])]
fu = float(tag.get('focalLengthU'))/64/binning_factor
fv = float(tag.get('focalLengthV'))/64/binning_factor
cu = float(tag.get('centerPointU'))/64/binning_factor \
- float(root.findall('cameras/camera/fov_video')[i].attrib.get('left'))
cv = float(tag.get('centerPointV'))/64/binning_factor \
- float(root.findall('cameras/camera/fov_video')[i].attrib.get('top'))
K += [np.array([fu, 0., cu, 0., fv, cv, 0., 0., 1.]).reshape(3,3)]
# Extrinsic parameters: rotation matrix and translation vector
for tag in root.findall('cameras/camera/transform'):
tx = float(tag.get('x'))/1000
ty = float(tag.get('y'))/1000
tz = float(tag.get('z'))/1000
r11 = float(tag.get('r11'))
r12 = float(tag.get('r12'))
r13 = float(tag.get('r13'))
r21 = float(tag.get('r21'))
r22 = float(tag.get('r22'))
r23 = float(tag.get('r23'))
r31 = float(tag.get('r31'))
r32 = float(tag.get('r32'))
r33 = float(tag.get('r33'))
# Rotation (by-column to by-line)
R += [np.array([r11, r21, r31, r12, r22, r32, r13, r23, r33]).reshape(3,3)]
T += [np.array([tx, ty, tz])]
# Cameras names by natural order
C_vid = [C[v] for v in vid_id]
Pose estimation test (#116) Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work! General: - Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend - The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version. - Added logging information to pose estimation - Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust) - Readme.md: added information about PoseEstimation with RTMLib - added poseEstimation to tests.py - created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well Json files: - the json files only saved one person, I made it save all the detected ones - tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge - took the save_to_openpose function out from the main function - minified the json files (they take less space when all spaces are removed) Detection results: - Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it. Others in Config.toml: - removed the "to_openpose" option, which is not needed - added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images'] - changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation) - added the flag: time_range_around_maxspeed in synchronization - automatically detect framerate from video, or set to 60 fps if we work from images (or give a value) - frame_range -> time_range - moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling) - removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead Others: - Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip) - scaling from motion instead of static pose (will have to study whether it's as good or not) - added logging to synchronization - Struggled quite a bit with continuous integration * Starting point of integrating RTMPose into Pose2Sim. (#111) * RTM_to_Open Convert format from RTMPose to OpenPose * rtm_intergrated * rtm_integrated * rtm_integrated * rtm_integrated * rtm * Delete build/lib/Pose2Sim directory * rtm * Delete build/lib/Pose2Sim directory * Delete onnxruntime-gpu * device = cpu * add pose folder * Update tests.py * added annotation * fix typo * Should work be still lots of tests to run. Detailed commit coming soon * intermediary commit * last checks before v0.9.0 * Update continuous-integration.yml * Update tests.py * replaced tabs with spaces * unittest issue * unittest typo * deactivated display for CI test of pose detection * Try to make continuous integration work * a * b * c * d * e * f * g * h * i * j * k * l --------- Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
C_vid_id = [C_vid.index(c) for c in sorted(C_vid, key=natural_sort_key)]
2023-07-19 17:37:20 +08:00
C_id = [vid_id[c] for c in C_vid_id]
C = [C[c] for c in C_id]
ret = [ret[c] for c in C_id]
S = [S[c] for c in C_id]
D = [D[c] for c in C_id]
K = [K[c] for c in C_id]
R = [R[c] for c in C_id]
T = [T[c] for c in C_id]
return C, S, D, K, R, T
def world_to_camera_persp(r, t):
2023-07-19 17:37:20 +08:00
'''
Converts rotation R and translation T
from Qualisys object centered perspective
to OpenCV camera centered perspective
and inversely.
Qc = RQ+T --> Q = R-1.Qc - R-1.T
'''
r = r.T
2024-01-25 18:02:58 +08:00
t = - r @ t
2023-07-19 17:37:20 +08:00
return r, t
def rotate_cam(r, t, ang_x=np.pi, ang_y=0, ang_z=0):
'''
Apply rotations around x, y, z in cameras coordinates
'''
rt_h = np.block([[r,t.reshape(3,1)], [np.zeros(3), 1 ]])
r_ax_x = np.array([1,0,0, 0,np.cos(ang_x),-np.sin(ang_x), 0,np.sin(ang_x),np.cos(ang_x)]).reshape(3,3)
r_ax_y = np.array([np.cos(ang_y),0,np.sin(ang_y), 0,1,0, -np.sin(ang_y),0,np.cos(ang_y)]).reshape(3,3)
r_ax_z = np.array([np.cos(ang_z),-np.sin(ang_z),0, np.sin(ang_z),np.cos(ang_z),0, 0,0,1]).reshape(3,3)
2024-01-25 18:02:58 +08:00
r_ax = r_ax_z @ r_ax_y @ r_ax_x
2023-07-19 17:37:20 +08:00
r_ax_h = np.block([[r_ax,np.zeros(3).reshape(3,1)], [np.zeros(3), 1]])
2024-01-25 18:02:58 +08:00
r_ax_h__rt_h = r_ax_h @ rt_h
2023-07-19 17:37:20 +08:00
r = r_ax_h__rt_h[:3,:3]
t = r_ax_h__rt_h[:3,3]
return r, t
def toml_write(toml_path, C, S, D, K, R, T):
'''
Writes calibration parameters to a .toml file.
'''
with open(os.path.join(toml_path), 'w+') as cal_f:
for c in range(len(C)):
cam=f'[cam_{c+1}]\n'
name = f'name = "{C[c]}"\n'
size = f'size = [ {S[c][0]}, {S[c][1]},]\n'
mat = f'matrix = [ [ {K[c][0,0]}, 0.0, {K[c][0,2]},], [ 0.0, {K[c][1,1]}, {K[c][1,2]},], [ 0.0, 0.0, 1.0,],]\n'
dist = f'distortions = [ {D[c][0]}, {D[c][1]}, {D[c][2]}, {D[c][3]},]\n'
rot = f'rotation = [ {R[c][0]}, {R[c][1]}, {R[c][2]},]\n'
tran = f'translation = [ {T[c][0]}, {T[c][1]}, {T[c][2]},]\n'
fish = f'fisheye = false\n\n'
cal_f.write(cam + name + size + mat + dist + rot + tran + fish)
meta = '[metadata]\nadjusted = false\nerror = 0.0\n'
cal_f.write(meta)
def calib_qca_to_toml_func(*args):
'''
Convert a Qualisys .qca.txt calibration file
to an OpenCV .toml calibration file
Usage:
import calib_qca_to_toml; calib_qca_to_toml.calib_qca_to_toml_func(r'<input_qca_file>')
2023-09-21 23:39:28 +08:00
OR calib_qca_to_toml -i input_qca_file
OR calib_qca_to_toml -i input_qca_file --binning_factor 2 -o output_toml_file
2023-07-19 17:37:20 +08:00
'''
try:
qca_path = args[0].get('input_file') # invoked with argparse
binning_factor = int(args[0]['binning_factor'])
if args[0]['output_file'] == None:
toml_path = qca_path.replace('.qca.txt', '.toml')
else:
toml_path = args[0]['output_file']
except:
qca_path = args[0] # invoked as a function
toml_path = qca_path.replace('.qca.txt', '.toml')
try:
binning_factor = int(args[1])
except:
binning_factor = 1
C, S, D, K, R, T = read_qca(qca_path, binning_factor)
RT = [world_to_camera_persp(r,t) for r, t in zip(R, T)]
2023-07-19 17:37:20 +08:00
R = [rt[0] for rt in RT]
T = [rt[1] for rt in RT]
RT = [rotate_cam(r, t, ang_x=np.pi, ang_y=0, ang_z=0) for r, t in zip(R, T)]
R = [rt[0] for rt in RT]
T = [rt[1] for rt in RT]
R = [np.array(cv2.Rodrigues(r)[0]).flatten() for r in R]
T = np.array(T)/1000
toml_write(toml_path, C, S, D, K, R, T)
print('Calibration file generated.\n')
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('-i', '--input_file', required = True, help='Qualisys .qca.txt input calibration file')
parser.add_argument('-b', '--binning_factor', required = False, default = 1, help='Binning factor if applied')
parser.add_argument('-o', '--output_file', required=False, help='OpenCV .toml output calibration file')
args = vars(parser.parse_args())
calib_qca_to_toml_func(args)