A Research Of Visual Intelligent Positioning System For The Elderly And Disabled Robot

The robot for the elderly and disabled is a kind of service robot for the elderly and the disabled people. It has the functions of helping the elderly, the disabled and entertainment. With the increase of the number of the elderly in the society,the demand for this robot has been increasing, and the research on this robot has been paid more attention.This thesis focuses on a visual intelligent positioning system that is applied to the immovable robot for the elderly and disable. We choose the Bumblebee2 to build our positioning system. After the deep understanding of the binocular stereo principle,we redeveloped this camera then it can realize the positioning of any target object and the positioning of the user's mouth.In order to positioning any object, we design a method using a laser pen in the user's head to point the object. This method is simple, convenient and practical and don't need to do objects' classification. The detection of the laser spot is smoothing the image which contains laser spot, do thresholding segmentation, erosion and dilation. The position of the laser spot in the image is obtained by the barycenter method. We can get the position's three-dimensional coordinates in the camera coordinate system.According to the Haar-like feature and Adaboost training method, the Haartraining algorithm provided by OpenCV is used to detect the face classifier and the human mouth classifier. The obtained classifier is used to detect the human face.In the face area, the position of the user's mouth in the image is detected, and the three-dimensional coordinates of the user's mouth in the camera coordinate system can be obtained by binocular stereoscopic vision.In the desktop PC, based on OpenCV and the API of the camera in the Microsoft Visual Studio 2008 programming environment programming to achieve the entire visual intelligence positioning system, the experimental results show that in the absence of strong light, The laser spot and user's mouth can be well recognized,the positioning accuracy of the positioning of the user's mouth of any other target in the 5mm range, to meet the needs of the project.