The invention discloses a binocular
image acquisition and preprocessing device and a method used for a
visual prosthesis. The device comprises left and right miniature cameras and an image preprocessing device. The left and right miniature cameras synchronously collect video information in real time and output the collected video information to the image preprocessing device for
image fusion, significant target extraction and other preprocessing. And then, a position, a distance and contour information of a significant target are further calculated. In the scheme, on one hand, the left and right miniature cameras are used to collect the image, and the image preprocessing device preprocesses the image; through using an
advantage of double cameras, a general orientation and distance information of the significant target in a visual field are calculated, and the above information is important for a blind person and is also ignored by other correlation patents; on the other hand, conversion and
information transmission are performed on significant target information in the visual field, and secondary information is neglected so that a calculated amount of a subsequent image coding module can be greatly reduced, and an operation speed of the
visual prosthesis is increased.