Design And Accomplishments Of Sound Navigation System Based On Digital Image Processing

The sense of hearing, as one of mankind’s most important sense organs, the meaning in the obstacle personage of eyesight is particularly important. Three professors of American Santa Barbara University of California branch and CarnegieMellon University introduced the auxiliary equipment of voice to one design for obstacle personage of eyesight together in1985. From then on, sound as a feedback in controling the navigation system start to develop its own way. And digital image processing technology has been developing into the research object of every field of social science nowadays from the initial figure compress aplication. Its way from the initial and simple picture to picture of process system develops into many kinds of application, appearing picture to voice, picture to large ways such as sense of touch in medical application.This thesis pays close attention to the research of changing visual information into the sound promptly.Mainly carry on systematic wave filter, edge detection, dawning the characteristic of the video information that CCD lens gathers,change into the sound according to the vectorial characteristic mapping and way that choose of form characteristic of edge, offering helps to the blind or visually impaired. Meanwhile, follow from quality analysis of picture, find one approach called DFD(depth from defocus),controllling the eletrical machinery drives the lens to achieve the goal of helping the one who has obstacle of eyesight to focus the scene flexibly automatically.The main research contents are as follows:(1) According to the structure and performance requirement of the Sound Based Guide Line System, make the overall design plan and frame structure of the thesis.(2) Based on the type of the lens and the image acquisition card, decided to capture the video and save the frame data by Directshow.(3) Combined the current focus method based on distance measurement, and the depth from focus method developed from Fibonacci search method and climb method, carry on an approach that gradually approaches to focus position called DFD.(4) Carry on the demand to measure the picture, design the improved scanning method according to the project TneVoice, scan the general environment and map to sound with parameter(shogen curve),and measure the edge of goal under the inactive state, carry on the particular sound and map to three kinds of common edge forms according to the characteristic of the form.(5) Successed in realizing sound mapping export in mobile and static mode. It has also realized the application of rough focusing.