This paper describes a new, occlusion-free, monocular three-dimensional vision system. A matrix of light beams (lasers, fiber optics, etc.), substantially parallel to the optic axis of the lens of a video camera, is projected onto a scene. The corresponding coordinates of the perspective image generated on the video-camera sensor, the focal length of the camera lens, and the lateral position of the projected beams of light are used to determine the "perspective depth" z* of the three-dimensional real image in the space between the lens and the image plane. Direct inverse perspective transformations are used to reconstruct the three-dimensional real-world scene. This system can lead to the development of three-dimensional real-image sensing devices for manufacturing, medical, and defense-related applications. If combined with existing technology, it has high potential for the development of three-dimensional television.