Device and method to detect an object and controlling a device to pick up the detected object

An object detecting device comprising:

• a memory;
• and
• a processor coupled to the memory, the processor is programmed to perform a process comprising:
• setting a plurality of external reference points used as information for estimating a position and an attitude of an object in external space of a model of the object, and setting an internal reference point used as information for determining whether the information for estimation is valid in internal space of the model;
• storing in the memory a table in which feature quantities on a local surface including a pair of a starting point and an endpoint that are sequentially selected from a point group located on a surface of the model are associated with a set of positions of the external reference points and the internal reference point with respect to the starting point;
• detecting a surface of the object existing in real space using a sensor;
• sequentially selecting a pair of the starting point and the endpoint from a sample point group located on the surface of the object existing in real space, and calculating feature quantities of the object on a local surface including the pair of the starting point and the endpoint;
• and
• acquiring, from the table, the set of positions associated with feature quantities matching the feature quantities of the object, transforming this set into a set of positions in the real space, and when the position of the internal reference point in the set of positions is outside the object, estimating the position and the attitude of the object with the positions of the external reference points in the set of positions excluded from the information for estimation.

An object detecting method comprising:

• setting a plurality of external reference points used as information for estimating a position and an attitude of an object in external space of a model of the object, and setting an internal reference point used as information for determining whether the information for estimation is valid in internal space of the model;
• storing a table in which feature quantities on a local surface including a pair of a starting point and an endpoint that are sequentially selected from a point group located on a surface of the model are associated with a set of positions of the external reference points and the internal reference point with respect to the starting point;
• detecting a surface of the object existing in real space using a sensor;
• sequentially selecting a pair of a starting point and an endpoint from a sample point group located on the surface of the object existing in real space, and calculating feature quantities of the object on a local surface including the pair of the starting point and the endpoint;
• and
• acquiring, from the table, the set of positions associated with feature quantities matching the feature quantities of the object, transforming this set into a set of positions in the real space, and when the position of the internal reference point in the set of positions is outside the object, estimating the position and the attitude of the object with the positions of the external reference points in the set of positions excluded from the information for estimation.