A primary travelling assistant system of bus detection and recognition for visually impaired people

It is a challenging task for many blind and visually impaired passengers travelling by bus. To assist visually impaired passengers to travel more independently, we design a computer vision-based system to detect and recognize bus information from images captured by a camera at a bus stop. Our system is able to notify the visually impaired people in speech the information of the coming bus, and detect the route number and other related information which is depicted in the form of text. In bus detection, histogram of the oriented gradient (HOG) descriptor is employed to extract the image-based features of bus facade. Cascade SVM model is applied to train a bus classifier to detect the existence of bus facade in sliding windows. In bus route number recognition, we design a text detection algorithm on the basis of layout analysis and text feature learning, and then recognize the text codes from detected text regions for audio notification. This algorithm is able to compute the image regions containing text information. Experimental results demonstrate the effectiveness of our proposed algorithm for bus detection and route number recognition.