def processor(cls, *args, **kwargs):
object_db = kwargs["object_db"]
object_id = kwargs.get("object_id", None)
submethod = kwargs["submethod"]
# db_models = Models(db_params, [ object_id ], method, submethod)
if object_id:
# TODO these should be loaded from the database?
json_K = kwargs["pipeline_params"]["K"]
json_D = kwargs["pipeline_params"]["D"]
imageWidth = kwargs["pipeline_params"]["imageWidth"]
imageHeight = kwargs["pipeline_params"]["imageHeight"]
document_ids = find_model_for_object(db_params_to_db(object_db.parameters()), object_id, model_type="mesh")
print document_ids
db_models = Documents(object_db, document_ids)
print "Found %d meshes:" % len(db_models)
else:
# TODO these should be loaded from the database?
json_K = []
json_D = []
imageWidth = 640
imageHeight = 480
db_models = []
return TransparentObjectsProcessor(
json_submethod=dict_to_cpp_json_str(submethod),
json_K=dict_to_cpp_json_str(json_K),
json_D=dict_to_cpp_json_str(json_D),
imageWidth=imageWidth,
imageHeight=imageHeight,
db_models=db_models,
)
def processor(cls, *args, **kwargs):
object_db = kwargs['object_db']
object_id = kwargs.get('object_id', None)
submethod = kwargs['submethod']
#db_models = Models(db_params, [ object_id ], method, submethod)
if object_id:
#TODO these should be loaded from the database?
json_K = kwargs['pipeline_params']['K']
json_D = kwargs['pipeline_params']['D']
imageWidth = kwargs['pipeline_params']['imageWidth']
imageHeight = kwargs['pipeline_params']['imageHeight']
document_ids = find_model_for_object(db_params_to_db(
object_db.parameters()),
object_id,
model_type='mesh')
print document_ids
db_models = Documents(object_db, document_ids)
print 'Found %d meshes:' % len(db_models)
else:
#TODO these should be loaded from the database?
json_K = []
json_D = []
imageWidth = 640
imageHeight = 480
db_models = []
return TransparentObjectsProcessor(
json_submethod=dict_to_cpp_json_str(submethod),
json_K=dict_to_cpp_json_str(json_K),
json_D=dict_to_cpp_json_str(json_D),
imageWidth=imageWidth,
imageHeight=imageHeight,
db_models=db_models)
def detector(self, *args, **kwargs):
visualize = kwargs.pop('visualize', False)
submethod = kwargs.pop('subtype')
parameters = kwargs.pop('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(), json_helper.dict_to_cpp_json_str(submethod))
return PBDDetector(object_db, object_ids, visualize, *args, **parameters['extra'])
def detector(self, *args, **kwargs):
#visualize = args.get('visualize', False)
submethod = kwargs.get('submethod')
parameters = kwargs.get('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(), json_helper.dict_to_cpp_json_str(submethod))
threshold = parameters.get('threshold', 90)
return ecto_linemod.Detector(model_documents=model_documents, db=object_db, threshold=threshold)
def detector(self, *args, **kwargs):
visualize = kwargs.pop('visualize', False)
submethod = kwargs.pop('submethod')
parameters = kwargs.pop('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(), json_helper.dict_to_cpp_json_str(submethod))
model_file = parameters['extra'].get('model_file')
return object_recognition_by_parts_cells.Detector(model_file=model_file, visualize=visualize)
def detector(self, *args, **kwargs):
visualize = kwargs.pop('visualize', False)
submethod = kwargs.pop('subtype')
parameters = kwargs.pop('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(),
json_helper.dict_to_cpp_json_str(submethod))
return PBDDetector(object_db, object_ids, visualize, *args,
**parameters['extra'])
def detector(self, *args, **kwargs):
visualize = kwargs.pop('visualize', False)
submethod = kwargs.pop('submethod')
parameters = kwargs.pop('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(), json_helper.dict_to_cpp_json_str(submethod))
registrationMaskFilename = parameters.get('registrationMaskFilename')
return transparent_objects_cells.Detector(model_documents=model_documents, object_db=object_db,
registrationMaskFilename=registrationMaskFilename, visualize=visualize)
def detector(self, *args, **kwargs):
#visualize = args.get('visualize', False)
submethod = kwargs.get('submethod')
parameters = kwargs.get('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(),
json_helper.dict_to_cpp_json_str(submethod))
threshold = parameters.get('threshold', 90)
return ecto_linemod.Detector(model_documents=model_documents,
db=object_db,
threshold=threshold)
def detector(self, *args, **kwargs):
visualize = kwargs.pop('visualize', False)
submethod = kwargs.pop('submethod')
parameters = kwargs.pop('parameters')
object_ids = parameters['object_ids']
object_db = ObjectDb(parameters['db'])
model_documents = Models(object_db, object_ids, self.type_name(),
json_helper.dict_to_cpp_json_str(submethod))
registrationMaskFilename = parameters.get('registrationMaskFilename')
return transparent_objects_cells.Detector(
model_documents=model_documents,
object_db=object_db,
registrationMaskFilename=registrationMaskFilename,
visualize=visualize)
options = parser.parse_args()
return options
if __name__ == '__main__':
options = parse_args()
plasm = ecto.Plasm()
#setup the input source, grayscale conversion
trainer = Trainer(path='/wg/stor2a/vrabaud/workspace/recognition_kitchen_groovy/build/buildspace/lib/object_recognition_reconstruction')
model_filler = ModelFiller()
#connect up the pose_est
plasm.connect(trainer['detector'] >> model_filler['detector'])
object_id = 'whoolite'
writer = ModelWriter(session_ids=list_to_cpp_json_str([]),
object_id=object_id, db_params=ObjectDbParameters({'type': 'CouchDB',
'root': 'http://bwl.willowgarage.com:5984',
'collection': 'object_recognition'}),
method=LinemodDetectionPipeline.type_name(),
json_submethod=dict_to_cpp_json_str({}),
json_params=dict_to_cpp_json_str({}))
plasm.connect(model_filler["db_document"] >> writer["db_document"])
options.niter = 1
run_plasm(options, plasm, locals=vars())
def incremental_model_builder(cls, *args, **kwargs):
submethod = kwargs.get('submethod')
return ecto_linemod.Trainer(json_submethod=dict_to_cpp_json_str(submethod))
if __name__ == '__main__':
options = parse_args()
plasm = ecto.Plasm()
db_params = ObjectDbParameters({'type': 'CouchDB', 'root': 'http://bwl.willowgarage.com:5984',
'collection': 'object_recognition'})
object_db = ObjectDb(db_params)
#setup the input source, grayscale conversion
from ecto_openni import VGA_RES, FPS_30
source = create_source('image_pipeline','OpenNISource',image_mode=VGA_RES,image_fps=FPS_30)
print LinemodDetectionPipeline.type_name()
object_ids = ['whoolite', 'tilex']
model_documents = Models(object_db, object_ids, LinemodDetectionPipeline.type_name(), json_helper.dict_to_cpp_json_str({}))
print len(model_documents)
detector = Detector(model_documents=model_documents, db=object_db, threshold=90)
#connect up the pose_est
plasm.connect(source['image'] >> detector['image'],
source['depth'] >> detector['depth']
)
plasm.connect(source['image'] >> imshow(name='source')[:])
if 0:
import ecto_ros
import ecto_ros.ecto_sensor_msgs
import sys
ecto_ros.init(sys.argv, "image_pub")
mat2image = ecto_ros.Mat2Image(frame_id='base', encoding='bgr8')
'collection': 'object_recognition'
})
object_db = ObjectDb(db_params)
#setup the input source, grayscale conversion
from ecto_openni import VGA_RES, FPS_30
source = create_source('image_pipeline',
'OpenNISource',
image_mode=VGA_RES,
image_fps=FPS_30)
print LinemodDetectionPipeline.type_name()
object_ids = ['whoolite', 'tilex']
model_documents = Models(object_db, object_ids,
LinemodDetectionPipeline.type_name(),
json_helper.dict_to_cpp_json_str({}))
print len(model_documents)
detector = Detector(model_documents=model_documents,
db=object_db,
threshold=90)
#connect up the pose_est
plasm.connect(source['image'] >> detector['image'],
source['depth'] >> detector['depth'])
plasm.connect(source['image'] >> imshow(name='source')[:])
if 0:
import ecto_ros
import ecto_ros.ecto_sensor_msgs
import sys
ecto_ros.init(sys.argv, "image_pub")
def incremental_model_builder(cls, *args, **kwargs):
submethod = kwargs.get('submethod')
return ecto_linemod.Trainer(
json_submethod=dict_to_cpp_json_str(submethod))
#setup the input source, grayscale conversion
trainer = Trainer(
path=
'/wg/stor2a/vrabaud/workspace/recognition_kitchen_groovy/build/buildspace/lib/object_recognition_reconstruction'
)
model_filler = ModelFiller()
#connect up the pose_est
plasm.connect(trainer['detector'] >> model_filler['detector'])
object_id = 'whoolite'
writer = ModelWriter(session_ids=list_to_cpp_json_str([]),
object_id=object_id,
db_params=ObjectDbParameters({
'type':
'CouchDB',
'root':
'http://bwl.willowgarage.com:5984',
'collection':
'object_recognition'
}),
method=LinemodDetectionPipeline.type_name(),
json_submethod=dict_to_cpp_json_str({}),
json_params=dict_to_cpp_json_str({}))
plasm.connect(model_filler["db_document"] >> writer["db_document"])
options.niter = 1
run_plasm(options, plasm, locals=vars())
