mi_recognition Class Reference
[Public interface]

Recognition class. More...

#include <mi_recognition.hh>

Inheritance diagram for mi_recognition:

[legend]Collaboration diagram for mi_recognition:

[legend]List of all members.

Public Member Functions
	mi_recognition (const mimas::mm_xml_reference< XalanElement > &configuration, bool _verbose)
	Constructor.
void	recognise (const mimas::mm_image< unsigned char > &image, double timestamp, mi_scene_ptr scene) const
	Perform recognition/tracking.
calibration_ptr	get_calibration (void)
	Access calibration object.
std::map< std::string, mi_object_type_ptr > &	get_object_types (void)
	Access list with types of micro-objects.
virtual bool	human_vision (const std::string &type, const mimas::mm_image< unsigned char > &image, double timestamp, double *angle, int *x, int *y) const
	Human vision.
Protected Member Functions
void	remove_objects (const std::string &method, mi_scene_ptr scene) const
	Remove some objects from the tracking-list.
void	reschedule (const std::string &method, mi_scene_ptr scene) const
	Remove some objects from the tracking-list.
void	take_best_matches (std::vector< match > &matches, double timestamp, mi_scene_ptr scene, vision_type method) const
	Take best recognition-results.
Protected Attributes
calibration_ptr	calib
	Calibration object.
std::map< std::string, recognition_method_ptr >	methods
	Initial recognition.
std::map< std::string, mi_object_type_ptr >	object_types
	General information about the different type of micro-objects.
bool	verbose

---

Detailed Description

Recognition class.

The recognition class hides the use of different recognition- and tracking-methods from the calling program.

This class is part of the public interface.

Author:

Jan Wedekind (j.wedekind@shu.ac.uk)

---

Constructor & Destructor Documentation

mi_recognition::mi_recognition (  const mimas::mm_xml_reference< XalanElement > &  configuration, bool  _verbose )

Constructor. Initialise the recognition algorithms and the database of micro-objects. Parameters: configuration  XML-tree with configuration. _verbose  Set to true, if you want to get verbose information from the vision-algorithms on stdout.

---

Member Function Documentation

calibration_ptr mi_recognition::get_calibration (  void   )  [inline]

Access calibration object. Get access to parameters of camera-calibration Returns: Pointer to camera-calibration object.

std::map< std::string, mi_object_type_ptr >& mi_recognition::get_object_types (  void   )  [inline]

Access list with types of micro-objects. A map is returned, which maps strings on pointers to mi_object_type objects. Returns: Reference to the map. See also: mi_object_type

virtual bool mi_recognition::human_vision (  const std::string &  type, const mimas::mm_image< unsigned char > &  image, double  timestamp, double *  angle, int *  x, int *  y )  const [virtual]

Human vision. Recognise objects using a human. If you want to recognise an object with human-vision, you need to override this method and modify the configuration-file (see f.e. gripper.xml). Parameters: type  Type of object to be recognised. image  Image to recognise object in. timestamp  Timestamp of image. angle  Angle of object. x  x-coordinate of object. y  y-coordinate of object. Returns: true, if recognition was successful. See also: human_vision_plugin

void mi_recognition::recognise (  const mimas::mm_image< unsigned char > &  image, double  timestamp, mi_scene_ptr  scene )  const

Perform recognition/tracking. The micro-objects known to the scene object are being tracked. Any new recognition-jobs, being specified in scene, are attempted. The model contained in scene is updated accordingly then. Parameters: image  Image to recognise/track micro-objects in. timestamp  Time of image-capture. scene  Object encapsulating the state of vision. See also: Units

void mi_recognition::remove_objects (  const std::string &  method, mi_scene_ptr  scene )  const [protected]

Remove some objects from the tracking-list. All entries of micro-objects being tracked with a specific vision-method are removed. Parameters: method  Tracking-method responsible for this object. scene  Scene-object with tracking-list.

void mi_recognition::reschedule (  const std::string &  method, mi_scene_ptr  scene )  const [protected]

Remove some objects from the tracking-list. All entries of micro-objects being tracked/recognised with a specific vision-method are rescheduled for recognition and removed from the tracking-list. Parameters: method  Tracking-method responsible for this object. scene  Scene-object with tracking-list and job-list.

void mi_recognition::take_best_matches (  std::vector< match > &  matches, double  timestamp, mi_scene_ptr  scene, vision_type  method )  const [protected]

Take best recognition-results. Parameters: matches  List of matches. timestamp  Timestamp of image. scene  Scene-object to be updated with new recognition-results. method  Type of vision-method (tracking or recognition).

---

The documentation for this class was generated from the following file:

• mi_recognition.hh

---