Depth interval estimates from motion parallax and binocular disparity beyond interaction space

Area of Science:

• Visual perception
• Depth perception
• Psychophysics

Background:

• Understanding how humans perceive depth at long distances is crucial for various applications, including autonomous navigation and virtual reality.
• Previous research has explored cues like stereopsis and motion parallax, but their effectiveness at extended ranges remains an area of investigation.

Purpose of the Study:

• To investigate the accuracy of static and dynamic binocular and monocular depth perception for objects beyond immediate interaction space.
• To determine the relative contributions of stereopsis and motion parallax in long-distance depth estimation.
• To examine the influence of environmental cues (ground plane, walls) on depth perception accuracy.

Main Methods:

• Participants estimated depths between pairs of LEDs presented in a dark railway tunnel, with distances ranging from 40m up to 288m.
• Depth estimates were collected using static and dynamic viewing conditions, employing both binocular and monocular vision.
• Environmental cues (tunnel walls and ground plane) were visible to assess their impact on scaled depth perception.

Main Results:

• Dynamic binocular viewing yielded the most veridical depth magnitude estimates, outperforming static binocular and dynamic monocular viewing.
• Static monocular viewing provided no significant depth perception.
• Both dynamic binocular and monocular depth estimates were scaled to the observation distance when environmental cues were present.

Conclusions:

• Motion parallax and stereopsis are both effective cues for perceiving depth at long distances.
• Motion parallax can significantly enhance the perceived stereoscopic depth, leading to more accurate estimations.
• The presence of environmental context aids in the accurate scaling of perceived depth over extended ranges.