The Binocular Camera Measuring About Vision Substitution Prosthesis

Artificial vision is a kind of high and new technology in the world and it is the application of modern electronics, biology and optical technology through different visual prosthesis which makes the blind regain vision. About135million people are visual disability in the global, in which there are about40to45million are the blindness. At present, there are about5million blindness in our country, so artificial vision research can get important social and economic benefits. Nowadays, artificial vision research approach include artificial vision prosthesis (the cortex prosthesis, optic nerve and retinal prosthesis) and vision substitution. Vision substitution refers to the development feeling (touch, hearing) alternative electronic auxiliary equipment, its research including signal extraction, signal conversion, signal recognition, etc. Our study is the preliminary research that the feeling replace of visual prosthesis of the mode of signal extraction and processing, which is the basic research of the vision substitution.In this article, the binocular stereo camera is used to extract the depth information of vision substitution. This thesis mainly focus on the hardware and software system structures, camera calibration, and stereo matching. Object to be tested depth information extraction on distance, which is base on the investigation of current research at home and abroad.1. According to parallel placed binocular stereo camera model, establishing the hardware platform of binocular stereo camera ranging system and configuring the software development environment.2. The plane of the calibration method template is applied to calibrate the binocular stereo camera, which is based on the OpenCV. According to this method, the radial and tangential distortion of lens must be fully considered, so the method is widely used in various scenarios. Through the experiment, extracting the calibration template angular point stably, calibrating inside and outside parameters of binocular stereo camera easily and quickly. The method has higher precision and good robustness provided by Bouguet’s calibration toolbox.3. In this paper the way of OpenCV stereo matching is used, the plane of the calibration method template is applied to calibrate the binocular stereo camera, which is based on the common stereo matching algorithm. In order to guarantee the correctness of the matching, reduce false matching rate, the introduction of a line constraint, image correction. The three-dimensional matching experiment gets ideal results, that establish the basic to the chapter below. 4. According to this article uses the parallel placed binocular stereo camera model, Extracting depth information by triangle measuring method of the measured object and getting the information of distance of the three objects (box, mineral spring water bottle, water cup), then measuring the depth of the actual information by steel tape. Finally, contrasting experiment data analyzing relative error. The results show that the system in a certain range is stable and accurate.