| CPC G06T 7/85 (2017.01) [G06T 2207/20081 (2013.01)] | 20 Claims |
|
1. A method comprising:
obtaining a plurality of simultaneous images of an object using a camera set, wherein:
the camera set comprises a first camera and a second camera, each obtaining one of the plurality of simultaneous images and collectively establishing a stereo axis of the camera set,
the camera set is a part of a camera unit further comprising one or more light units, a support structure maintaining stable relative positions of the camera set and light units, and a camera-unit actuator,
the one or more light units normalize major variations of ambient light conditions, thereby reducing camera integration times and providing consistent image characteristics to benefit machine learning performance, and
the one or more light units illuminate for a duration between 100-200 microseconds during which the camera set obtains the plurality of simultaneous images, thereby producing consistent images for constructing a three-dimensional (3D) model;
constructing a three-dimensional (3D) model from the plurality of simultaneous images using a control unit of the camera unit; and
when fidelity of the 3D model is insufficient,
(a) reconfiguring the camera unit using the camera-unit actuator such that the stereo axis of the camera unit has a different orientation relative to the object,
(b) obtaining a plurality of new simultaneous images of the object, and
(c) updating the 3D model using the plurality of new simultaneous images with the stereo axis of the camera unit being in a different orientation relative to the object.
|