400bd60 has a few things that should be useful for testing
it creates test folders automatically.
the typical directory where test data is saved is configurable via a ROS parameter called, /odcl/odcldir
by default (i.e. if unset) the test data is saved to ~/odcldata/
when the node is run, it will create a folder in ~/odcldata/ with today's date. So each day of testing will have its own date, making test data easier to manage
two folders are made, raw, and crop, where images are stored
a CSV file of each "run" of the odcl pipeline is also created. following An's idea each run has a prefix (01, 02...)
The raw image is saved with a unique identifier in /raw/
The cropped image of each target is saved with a unique identifier in /crop/
9e9360f has a few additions
Each found target is given a unique identifier, which is the date, run, frame, target.
along with the 5 odcl parameters (lat lon scolor lcolor shape) we also have several new parameters and metadata
run number, frame number, target number, and target ID
time taken to inference
class confidence score
5ed9350
Whenever an image is captured as part of the pipeline, the exact capture time is recorded and saved
400bd60 has a few things that should be useful for testing
/odcl/odcldir
~/odcldata/
~/odcldata/
with today's date. So each day of testing will have its own date, making test data easier to manageraw
, andcrop
, where images are storedThe raw image is saved with a unique identifier in
/raw/
The cropped image of each target is saved with a unique identifier in/crop/
9e9360f has a few additions
Each found target is given a unique identifier, which is the date, run, frame, target.
5ed9350
Whenever an image is captured as part of the pipeline, the exact capture time is recorded and saved