Configuration marker design and detection for instrument tracking

A robotic surgery system, comprising:

• a robotic surgical tool having a plurality of primitive features having know positions on the tool, and wherein each feature comprises a spherical reflective surface;
• a light source oriented to transmit illumination light within a patient body;
• a stereo image capture device adjacent the illumination source so that the illumination light reflected from the primitive features toward the image capture device substantially aligns with a spherical centers of the spherical surfaces;
• and
• a processor coupled with the image capture device and configured for determining a position for the tool by processing stereo images so as to locate the spherical centers of imaged primitive features by using the reflected light.

An object tracking system, comprising:

• an object having a plurality of primitive features with each primitive feature comprising a spherical reflective surface;
• a light source oriented to transmit illumination light toward the object;
• a stereo image capture device for capturing stereo images of said object, the image device being disposed adjacent the illumination source so that illumination light reflected from a plurality of primitive features toward the image capture device substantially aligns with spherical centers of the spherical surfaces, the image device outputting image data for said stereo images;
• and
• a processor coupled with the image capture device and configured to process said image data so as to:
• determine three-dimensional position data for three or more of the imaged primitive features;
• and
• determine a position for the tool by processing the three-dimensional position data.

A method for estimating the pose of a surgical tool comprising three or more substantially corner-less primitive features having known positions on the tool, the method comprising:

• using a stereoscopic endoscope to capture stereo images of three or more of the primitive features, the stereo images including a first image and a second image;
• extracting at least three of the primitive features from the first image;
• extracting at least three of the primitive features from the second image;
• determining correspondences between extracted primitive features by using image signatures;
• using the determined correspondences to determine three-dimension positions for at least three of the primitive features;
• identifying a pattern in the extracted primitive features that corresponds to a pattern of the tool primitive features;
• and
• estimating a pose for the surgical tool by using the identified pattern and the determined three-dimensional positions.