Methods And Technology On Books Automatically Picking By Robot In The Library

With the progress of science and technology, the automation technology developed very fast. However, the management of library didn’t go with the time. The works of finding, acquiring, putting back and moving the books are still relying on human hands. Process of these works are cumbersome, inefficient and fallibility. In order to solve these problems, we propose an idea to use the intelligent robots with the task of acquiring books in the library. The solution includes precise location, autonomous navigation, books identification and books accessing, etc. In this paper, we focus on several key points of the solution as follows:Firstly, in view of the internal structure of the library, we choose the bookcases for environment landmarks, with which we construct the feature map. The feature map records all coordinates of landmarks, which lays the foundation for precise positioning and autonomous robot navigation.Secondly, two cases of landmarks identified are discussed. On the side of books shelves, laser sensor can obtain regular data, which can be clustered into several geometry lines. The lines presented for the landmarks can be easily recognized. While laser sensor data obtained from the front of books shelves is more complicated, it is difficult to recognize the landmarks using single laser sensor. Fortunately, the front of books shelves get plenty optical character image regions, by combining the use of laser sensor and vision sensor can the robot detect and locate landmarks.Then robot finishes accurate positioning by using landmarks which have been detected previously and completes the navigation by tracking the landmarks. While moving towards the target, the extended Kalman filter is used for the prediction and correction of the robot’s position and attitude, with which can we reduce the impact of systematic errors caused by the robot navigation.Finally, a method of particle clustering algorithm is proposed for image segmentation. We use a kind of improved particle swarm optimization algorithm to extract the book’s tagged image from the image of front bookcase and remove noise points by using k-Means clustering algorithm. While the character images are arranged closely and edge-rich, we can easily extract character images from book’s tagged image.In summary, researches of robot map building, landmarks recognition, robot navigation, books’ tags recognition provide a theoretical and practical basis for the intelligent robot to pick up the books automatically in the library.