Research On Indoor Guide System Based On Visual-Audiory Conversion

As the world's population ages and the number of people who are blind and visually impaired continues to rise,researchers around the world have been exploring ways to use other human organs to perceive their surroundings.Currently,the relevant guide blind system mainly uses ultrasonic wave,laser,camera and other systems to collect information,while the camera can collect richer information of the surrounding environment,and the image processing is more convenient.The information output mode of the guide system is generally to receive information by touch or hearing,which has better perception ability than touch.So in this paper,based on visual and auditory convert indoor guide system is studied,the system through the depth camera Xtion gathers the image of the surrounding environment on the depth of collection of images for image preprocessing,image restoration,image segmentation,such as operation,get obstacles in image information,the information through the image-mapping scheme is converted to sound parameters,combined with the virtual stereo audio,finally played through headphones to the blind,to help it walk in indoor security.The main research of this paper includes:1.Aiming at the problem that the hollow area in the depth image obtained by the depth camera will affect the subsequent image processing and user judgment.This paper proposes an improved guidance repair method based on the weight function,which is compared with the traditional depth repair method.The quantitative analysis of the three evaluation indicators shows that the improved method of this paper effectively reduces the void area in the image.2.Due to obstacles in image information and the background is exist at the same time,information transmission and guide system only needs to be obstacles to the user,and the background information in the image may interfere with the blind users on the surrounding environment,so on the obstacles and background in image segmentation,information extract clean complete obstruction.This paper the characteristics of the human visual attention,this paper proposes a figure significantly improved segmentation method based on super division of pixels,this method will super pixel segment and extract significant figure by attention model,and compared with traditional image segmentation method for experiments and several objective evaluation indicators and the results show that this method has a more complete segmentation effect.3.In order to convey the obstacle information obtained above to blind users,an imagesound mapping scheme based on virtual stereo is proposed.By combining the parameters of effective obstacle information and virtual stereo,audio data of surrounding environment information is obtained.This scheme is evaluated and analyzed with several traditional imagesound mapping schemes,and the evaluation results show that the proposed scheme has higher obstacle resolution and satisfaction.4.Finally,the indoor guide system studied in this paper is compared with the two guide systems.The test results show that the guide system in this paper performs better in terms of obstacle resolution and user satisfaction.