Indoor Mobile Robot Path Planning Based on Visual
|Course||Pattern Recognition and Intelligent Systems|
|Keywords||region segmentation modeling grid method genetic algorithm path planning|
One task of the robot navigation is avoiding obstacles, so the robot in the process of walking, to be able to distinguish between the feasible region and obstacle area. With the development of computer and image processing technology swift and violent development, robot vision navigation are constantly being developed. At present, most mobile robot navigation based on the involvement of visual or visual navigation technology.In this paper, the first is to solve the problem is through the robot visual access to the surrounding environment. After image preprocessing, remove images of random noise interference, enhance the image of the useful information, the interception of images in a portion of the ground image, obtained in the RGB color space under the ground characteristic value, including gray ground, three color components of the mean and the standard deviation of the values of these features, the entire image the pixels are detected one by one, segmentation of the feasible region and obstacle area, and the segmented region two value processing, and then the image expansion and other post-processing, to facilitate follow-up work on robot vision environment grid modeling.In this paper second problems is to be solved on the forward line of path planning of robot. The first is processed to the value of two images of the grid method in modeling, grid map identifies the feasible region and obstacle area, then the robot current view on regional application of genetic algorithm to optimal planning of robot local path, every step, will in the field were detected, and then planning the current view the local optimal path, in order to achieve global optimal path planning.