Study On Blind Image Segmentation Technology Based On Computer Vision

Nowadays,it is very incovenient for blind people to go outside.Blind road,as an important means to assist blind people has not been fully exploited.Therefore,how to effectively use the blind road information to facilitate the blind travel has become the focus of researchers.As part of the blind system,the study of blind segmentation is of great significance.However,the existing blind separation algorithm has shortage such as poor resolution,single processing,and easy to be affected by light,shadow and other external environment etc.Aiming at these shortcomings,this paper presents a new algorthm based on computer vision.According to different segmentation algorithms have different effects on different types of blind paths,this paper divides the blind division into two parts: texture blind division and color blind division.Texture blind is a texture image,and commonly used texture image segmentation algorithm are watershed algorithm and graph theory algorithm,but these algorithms are subject to light and noise and other external factors,thus easy to cause misclassification.In view of these shortcomings,this paper presents a blind algorithm based on BBO-KFCM.Firstly,the texture feature of the blind channel is extracted by the structure method.Then,the KFCM clustering algorithm is chosen to segment the image.Finally,the BBO is used to reduce the sensitivity of the algorithm to noise.In the color blind division,the segmentation algorithm based on the region algorithm and the threshold algorithm based on the threshold algorithm show a weaker segmentation effect and robustness.In this paper,proposed a blind algorithm for BBO-FCM based on HSI color space.Firstly,the image is transformed from RGB color space to HSI color space,and the European distance of HSI color space is improved.Secondly,the improved FCM algorithm and BBO algorithm are combined to optimize the FCM algorithm by using the BBO algorithm,to enhance the anti-noise ability of FCM.In this paper,the blind path segmentation algorithm is studied by computer vision method.In view of the shortcomings of the existing blind path segmentation algorithm,blind path is divided into texture blindness and color blindness,and the improved algorithm is proposed accordingly.Experimental results show that the proposed algorithm can effectively divide the blind area from the pedestrian area for different types of blind paths,and is less susceptible to illumination and noise,thus reducing the error rate.