Monocular cued detection of three-dimensional structures from depth images

A method for detection of obstacles, the method performed by a system mountable in a vehicle, wherein the system includes a processor operatively connectible to a camera, the method comprising:

• capturing a plurality of two-dimensional image frames in a field of view of the camera;
• detecting, in the image frames, a two-dimensional line associated with a candidate three-dimensional object in an environment of the vehicle using monocular information;
• portioning the image frames around the line to produce low-resolution image portions including the line and high-resolution image portions not including the line;
• computing depth values for points around the line in the low-resolution image portions, including depth values related to pixels on the line;
• and
• identifying, using a depth map, the object associated with the line.

A system for detection of obstacles, the system mountable in a host vehicle, the system including a processor operatively connectible to a camera, wherein the processor is configured to:

• capture a plurality of two-dimensional image frames in a field of view of the camera;
• detect, in the image frames, a two-dimensional line associated with a candidate three-dimensional object in an environment of the vehicle using monocular information;
• portion the image frames around the line to produce low-resolution image portions including the line and high-resolution image portions not including the line;
• compute depth values for points around the line in the low-resolution image portions, including depth values related to pixels on the line;
• and
• identify, using a depth map, the object associated with the line.

A non-transitory computer-readable storage medium storing instructions for detection of obstacles, the instructions causing one or more processors to:

• capture a plurality of two-dimensional image frames in a field of view of the camera;
• detect, in the image frames, a two-dimensional line associated with a candidate three-dimensional object in an environment of the vehicle using monocular information;
• portion the image frames around the line to produce low-resolution image portions including the line and high-resolution image portions not including the line;
• compute depth values for points around the line in the low-resolution image portions, including depth values related to pixels on the line;
• and
• identify, using a depth map, the object associated with the line.