Abstract: Using mirrors to capture, by a camera of a robot, images that capture portions of an environment from multiple vantages. In various implementations, multiple images, that each capture a portion of an environment from a different vantage, may be captured by a camera sensor of a robot via the adjustment of one or more mirrors viewable by the camera sensor—and the multiple images may be captured independent of locomotion of the robot and/or independent of adjusting a camera sensor pose of that camera sensor. One or more processors associated with the robot may utilize the multiple images to determine one or more features about those portions of the environment captured from different vantages in the images.

Abstract: An example method includes receiving a plurality of detected depth points indicative of depths of at least one surface and determining a projection of the detected depth points onto a plane. The method may also include identifying a plurality of first detected points, where a first detected point comprises a first point at a particular location of the plane in the projection. The method may also include storing digital entries corresponding to points located within a threshold buffer from one of the first detected points relative to the plane. The method may additionally include determining values for the digital entries, where a value for a digital entry corresponding to a particular point comprises an accumulation of distances from detected depth points that cross the particular point. The method may further include determining a digital height map representative of heights of the at least one surface relative to the plane.