| As the visual system can provide accurate and complete environmental information, autonomous mobile robot based on visual navigation has become a hotspot in the research field of robot. Setting up the visual system in autonomous mobile robot navigation as the main line, a survey of self-localization, obstacle detection and local path planning in the navigation task is carried out in this thesis.The main research work and achievements of the thesis are as follow:Firstly,a self-localization method based on skirt lines in the corridor environment is proposed, in which skirt lines on both sides of the corridor are considered as natural sign of robot self-localization. The feature extraction of skirt lines by Hough transform is researched, and the location of robot is estimated by the status of skirt lines in the image. Experiments have proved that proposed method is able to detect the current status, direction and yawed distance from the road of the robot, and it can adjust the robot correspondingly on real-time.Moreover, a color image adaptive segmentation algorithm based on HSI space is suggested, whose application in detecting mobile area is investigated in the thesis. The color space of image is transformed to HSI space, and the region is divided into movable and barriers by comparing each pixel in the image with the average color and saturation in the reference surface pixels. The experiments results show that the method can effectively overcome the environmental change interference, such as light intensity. In addition, the algorithm has the advantage of adaptability, regional segmentation accuracy.Furthermore, a study of the robot local path planning methods and the grid-based dynamic path planning method under special environment is carried out in the thesis. The environment is converted to a grid map to recognize obstacles. There is experimental evidence to suggest that it is effective, simple and accurate when it is applied to avoid obstacles. |
PDF Full Text Request