May I know how can blender save each frame info to a txt file when rendering a video using python
the code is:
bpy.context.scene.frame_start = 1
bpy.context.scene.frame_end = 13
bpy.ops.render.render(write_still = False)
# save frame information into txt file
the running cmd is:
blender -b env.blend -o ./ --python run_blender.py -a
When render process is over, how can I save each frame info into txt
Related
I'm following the tutorial on using your own template images to do object 3D pose tracking, but I'm trying to get it working on Ubuntu 20.04 with a live webcam stream.
I was able to successfully make my index .pb file with extracted KNIFT features from my custom images.
It seems the next thing to do is load the provided template matching graph (in mediapipe/graphs/template_matching/template_matching_desktop.pbtxt) (replacing the index_proto_filename of the BoxDetectorCalculator with my own index file), and run it on a video input stream to track my custom object.
I was hoping that would be easiest to do in python, but am running into dependency problems.
(I installed mediapipe python with pip3 install mediapipe)
First, I couldn't find how to directly load a .pbtxt file as a graph in the mediapipe python API, but that's ok. I just load the text it contains and use that.
template_matching_graph_filepath=os.path.abspath("~/mediapipe/mediapipe/graphs/template_matching/template_matching_desktop.pbtxt")
graph = mp.CalculatorGraph(graph_config=open(template_matching_graph_filepath).read())
But I get missing calculator targets.
No registered object with name: OpenCvVideoDecoderCalculator; Unable to find Calculator "OpenCvVideoDecoderCalculator"
or
[libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/text_format.cc:309] Error parsing text-format mediapipe.CalculatorGraphConfig: 54:70: Could not find type "type.googleapis.com/mediapipe.TfLiteInferenceCalculatorOptions" stored in google.protobuf.Any.
It seems similar to this troubleshooting case but, since I'm not trying to compile an application, I'm not sure how to link in the missing calculators.
How to I make the mediapipe python API aware of these graphs?
UPDATE:
I made decent progress by adding the graphs that the template_matching depends on to the cc_library deps of the mediapipe/python/BUILD file
cc_library(
name = "builtin_calculators",
deps = [
"//mediapipe/calculators/image:feature_detector_calculator",
"//mediapipe/calculators/image:image_properties_calculator",
"//mediapipe/calculators/video:opencv_video_decoder_calculator",
"//mediapipe/calculators/video:opencv_video_encoder_calculator",
"//mediapipe/calculators/video:box_detector_calculator",
"//mediapipe/calculators/tflite:tflite_inference_calculator",
"//mediapipe/calculators/tflite:tflite_tensors_to_floats_calculator",
"//mediapipe/calculators/util:timed_box_list_id_to_label_calculator",
"//mediapipe/calculators/util:timed_box_list_to_render_data_calculator",
"//mediapipe/calculators/util:landmarks_to_render_data_calculator",
"//mediapipe/calculators/util:annotation_overlay_calculator",
...
I also modified solution_base.py so it knows about BoxDetector's options.
from mediapipe.calculators.video import box_detector_calculator_pb2
...
CALCULATOR_TO_OPTIONS = {
'BoxDetectorCalculator':
box_detector_calculator_pb2
.BoxDetectorCalculatorOptions,
Then I rebuilt and installed mediapipe python from source with:
~/mediapipe$ python3 setup.py install --link-opencv
Then I was able to make my own class derived from SolutionBase
from mediapipe.python.solution_base import SolutionBase
class ObjectTracker(SolutionBase):
"""Process a video stream and output a video with edges of templates highlighted."""
def __init__(self,
object_knift_index_file_path):
super().__init__(binary_graph_path=object_pose_estimation_binary_file_path,
calculator_params={"BoxDetector.index_proto_filename": object_knift_index_file_path},
)
def process(self, image: np.ndarray) -> NamedTuple:
return super().process(input_data={'input_video':image})
ot = ObjectTracker(object_knift_index_file_path="/path/to/my/object_knift_index.pb")
Finally, I process a video frame from a cv2.VideoCapture
cv_video = cv2.VideoCapture(0)
result, frame = cv_video.read()
input_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
res = ot.process(image=input_frame)
So close! But I run into this error which I just don't know what to do with.
/usr/local/lib/python3.8/dist-packages/mediapipe/python/solution_base.py in process(self, input_data)
326 if data.shape[2] != RGB_CHANNELS:
327 raise ValueError('Input image must contain three channel rgb data.')
--> 328 self._graph.add_packet_to_input_stream(
329 stream=stream_name,
330 packet=self._make_packet(input_stream_type,
RuntimeError: Graph has errors:
Calculator::Open() for node "BoxDetector" failed: ; Error while reading file: /usr/local/lib/python3.8/dist-packages/
Looks like CalculatorNode::OpenNode() is trying to open the python API install path as a file. Maybe it has to do with the default_context. I have no idea where to go from here. :(
I am using YOLO for model training. I want to crop the detected object.
For Darknet repository am using is: https://github.com/AlexeyAB/darknet/
For Detection and storing output coordinates in a text file am using this:
!./darknet detector test data_for_colab/obj.data data_for_colab/yolov3-tiny-obj.cfg yolov3-tiny-obj_10000.weights -dont_show -ext_output < TEST.txt > result.txt
Result.jpg
Considering in the TEST.txt file you have details as the sample image.
You can use re module of python for text pattern detection, ie your "class_name".
Parsing the .txt file
import re
path='/content/darknet/result.txt'
myfile=open(path,'r')
lines=myfile.readlines()
pattern= "class_name"
for line in lines:
if re.search(pattern,line):
Cord_Raw=line
Cord=Cord_Raw.split("(")[1].split(")")[0].split(" ")
Now we will get the coordinates in a list.
Coordinate calculation
x_min=int(Cord[1])
x_max=x_min + int(Cord[5])
y_min=int(Cord[3])
y_max=y_min+ int(Cord[7])
Cropping from the actual image
import cv2
img = cv2.imread("Image.jpg")
crop_img = img[y_min:y_max, x_min:x_max]
cv2.imwrite("Object.jpg",crop_img)
I'm currently working on a mobilenet pre-trained network which I would like to re-train with a dataset which contains png images.
I call the retrain script as follow :
python scripts/retrain.py
--bottleneck_dir=tf_files/bottlenecks
--how_many_training_steps=200
--model_dir=tf_files/models/
--summaries_dir=tf_files/training_summaries/"mobilenet_0.50_224"
--output_graph=tf_files/retrained_graph.pb
--output_labels=tf_files/retrained_labels.txt
--architecture mobilenet_0.50_224
--image_dir=tf_files/data
It seems like the images needs to be jpg, is it any way to work with png images instead ?
Can confirm it doesn't work with png files. I have, however, written a bash script that when placed in the same directory as the subclasses of the dataset can convert the images to jpg.
first you need to install imagemagick package by:
sudo apt-get install imagemagick
then you can run this script:
#!/bin/bash
for d in */ ; do
cd "$d"
for p in * ; do
IFS='.' read -r -a array <<< "$p"
convert "$p" "${array[0]}".jpg
done
cd ..
done
edit:
retrain.py does have a list with valid extensions (line 151):
extensions = ['jpg', 'jpeg', 'JPG', 'JPEG']
I didn't try to add 'png' to the list though
I am looking for a way to print 3D pdf from the results ABAQUS/Viewer. This will make it easy to communicate the results with others who are interested in the results of simulation but do not have access to ABAQUS.
The best way is to export a vrml file and convert it using Tetra4D or pdf3D and Adobe Acrobat professional. The 3D pdfs can look very good. However, the commercial software would cost over £800 per year. I did create a Python script to create a 3D pdf directly from Abaqus/CAE & Viewer which uses 2 open source tools: 1) Meshlab (http://www.meshlab.net/) to create a U3D file, 2) MiKTeX (https://miktex.org/) to convert the U3D file into a pdf. The output is not as polished as Tetra4D but it works. I have not tried this with the latest version of Meshlab. Just run this script from Abaqus/CAE or Abaqus/Viewer.
# Abaqus CAE/Viewer Python Script to create a 3D pdf directly from Abaqus/CAE or Abaqus/Viewer.
# You must first install meshlab (meshlabserver.exe)and MiKTeX (pdflatex.exe)
# Edit this script to reflect the installed locations of meshlabserver.exe and pdflatex.exe
# It will export a stl or obj file the mesh of current viewport and convert into 3D pdf
# Or run in Abaqus/viewer and it will create a VRML file and convert to 3D pdf.
# If contours are displayed in Abaqus Viewer, then it will create a contour 3D pdf
from abaqus import *
from abaqusConstants import *
from viewerModules import *
import os
import subprocess
import sys
# -----------------------------------------------------------------------------
pdfName='try'
meshlab_path="C:/Program Files/VCG/MeshLab/meshlabserver.exe"
pdfLatex_path="C:/Program Files (x86)/MiKTeX 2.9/miktex/bin/pdflatex.exe"
# -----------------------------------------------------------------------------
currView=session.viewports[session.currentViewportName]
try: # for Abaqus Viewer
cOdbD=currView.odbDisplay
odb = session.odbs[cOdbD.name]
name=odb.name.split(r'/')[-1].replace('.odb','')
module='Vis'
except: # Abaqus CAE
#name=currView.displayedObject.modelName
import stlExport_kernel
name = repr(currView.displayedObject).split('[')[-1].split(']')[0][1:-1] # allows for either main or visulation modules
module='CAE'
print module
if module=='CAE':
#All instances must be meshed
cOdbD=None
try:
ext='.stl'
stlExport_kernel.STLExport(moduleName='Assembly', stlFileName=pdfName + ext, stlFileType='BINARY')
except:
try:
ext='.obj'
session.writeOBJFile(fileName=os.path.join(directory,pdfName + ext), canvasObjects= (currView, ))
except:
print 'Either your assembly is not fully meshed or something else'
directory=(os.getcwd())
else: # Abaqus/Viewer
if cOdbD.viewCut:
session.graphicsOptions.setValues(antiAlias=OFF) # Better with anti aliasing off
odb = session.odbs[cOdbD.name]
directory=odb.path.replace(odb.path.split('/')[-1],'').replace('/','\\')
# Turn off most of the stuff in the viewport
currView.viewportAnnotationOptions.setValues(triad=OFF,
legend=OFF, title=OFF, state=OFF, annotations=OFF, compass=OFF)
ext='.wrl'
session.writeVrmlFile(fileName=os.path.join(directory,pdfName + ext),
compression=0, canvasObjects= (currView, ))
pdfFilePath=os.path.join(directory,pdfName+'-out.pdf')
if os.path.isfile(pdfFilePath):
os.remove(pdfFilePath)
#Check file was deleted
if os.path.isfile(pdfFilePath):
print "Aborted because pdf file of same name cant be deleted. Please close programs which it might be open in"
1/0 #a dodgy way to exit program
# Invoke meshlab to convert to a .u3d file
if cOdbD: #If in Abaqus/viewer
if 'CONTOURS' in repr(cOdbD.display.plotState[0]): # If contours are displayed. Output contoured pdf
p=subprocess.Popen([meshlab_path,'-i',pdfName + ext, '-o',pdfName + '.u3d','-m','vc']) #'vn fn fc vt'
else:
p=subprocess.Popen([meshlab_path,'-i',pdfName + ext, '-o',pdfName + '.u3d'])
else:
p=subprocess.Popen([meshlab_path,'-i',pdfName + ext, '-o',pdfName + '.u3d'])
p.communicate() # Wait for meshlab to finish
file_fullPathName=os.path.join(directory, pdfName + '.tex')
#Read the .tex file which meshlab has just created
with open(file_fullPathName, 'r') as texFile:
lines = texFile.read()
#Edit the .tex file
lines=lines.replace("\usepackage[3D]{movie15}","\\usepackage[3D]{movie15}\n\\usepackage[margin=-2.2in]{geometry}")
if cOdbD:
if 'CONTOURS' in repr(cOdbD.display.plotState[0]):
lines=lines.replace("3Dlights=CAD,","3Dlights=CAD,\n\t3Drender=SolidWireframe,")
lines=lines.replace("\n\end{document}","{---------------------------------------------------------------------------------Click above! MB1 - rotate, MB2 wheel or MB3 - zoom, Ctrl-MB1 - pan--------------}\n\\end{document}")
file_fullPathName=os.path.join(directory, pdfName + '-out.tex')
with open(file_fullPathName, "w") as outp:
outp.write(lines)
p=subprocess.Popen([
pdfLatex_path,
pdfName + '-out.tex',
])
p.communicate()
print 'Conversion to pdf complete'
print file_fullPathName
The simplest way of printing the Abaqus *.odb results are using Tecplot 360 which is read the Abaqus *.odb files and you can get the *.tif and *.png results with any resolutions and you can also rotate the model in 3D and change the fonts and all the things you need.
I use gnuplot with epslatex option to generate figure files for plotting purposes (like here). Via this method you get 2 files corresponding to same figure, one tex file and one eps file. The figure information is in eps file and font information is in tex file. So my question is this :
Can I combine both font information and figure content to a single file like pdf / eps file ?
UPDATE : OK I forgot to mention one thing. Off course set terminal postscript eps will give me eps outputs, but it will not embed latex symbols in the plot as labels etc.
So I found a method which I got from Christoph's comment. Set terminal like set terminal epslatex 8 standalone and then finally after plotting do something like below:
set terminal epslatex color standalone
set output "file.tex"
set xrange [1:500]
set ylabel "Variance (\\AA\\textsuperscript{2})" # angstoms
set mxtics 4
plot "version1.dat" using 1:3 with linespoints pointinterval -5 pt 10 lt 1 lw 3 title 'label1' , \
"version1.dat" using 1:2 with linespoints pointinterval -5 pt 6 lt -1 lw 3 title 'label2';
unset output
# And now the important part (combine info to single file) :
set output # finish the current output file
system('latex file.tex && dvips file.dvi && ps2pdf file.ps')
system('mv file.ps file.eps')
unset terminal
reset
These steps do output tex file which is converted to dvi and ps file. And finally you rename the postscript file to eps. Now you have figure information and tex symbol information in single file. This eps file is accepted by latex files.
OK now why this works : Sorry I don't know the entire technical details. But this is working fine with me.