Methods and systems for detecting obstacles for a visual prosthesis

Abstract

Systems and methods for detecting and displaying obstacles to visually impaired persons are disclosed. Images from a captured scene are analyzed to detect obstacles. A V-Disparity algorithm is employed to detect the ground plane and remove it from the images, thereby enhancing the capacity of retinal implants to display only obstacles. The system includes cameras to capture images; a processor with memory to detect obstacles in the captured images, to remove a ground plane, and to calculate a map of pixels representing the obstacles; and a retinal implant configured to receive the map of pixels and display the obstacles to a system user by electrically stimulating retinal nerves. The method includes synchronizing captured images and calculating a disparity map between the captured images to detect obstacles and remove a ground plane from processed images. A map of pixels representing the obstacle is transmitted to a retinal implant. Depth information may derive from stereo cameras, a time of flight camera or a structured light camera.