Mobile Robot Vision Technology-based Target Search And Tracking

With the development of robot vision technology, how to make the robot monitoring automatically has become a hotpot in the research of robot vision. Right now, monitor systems have been widely used in lots of fields, such as in Industrial, in Mining, in Super Market, in Intelligent Traffic, and others. In the computer monitor systems, we usually use image acquisition devices, such as camera, to obtain images. Due to the fixing of the camera, the fields of the vision are limited. Furthermore, the target may move out of the vision field. So, it is necessary to keep the object target in the vision. Actually, we can use several cameras to solve the vision field problems. If so, it will increase the cost of the whole system and require a better calculation unit for image process.In our research, in order to shorten the search time, we proposed a search order based on the uniform distribution if the position of the target object; we also gave the optimal rotation angle, which makes PaPeRo rotate as much as it can, meanwhile it can still see the entire target object; by learn from AR technique, we implemented the target object matching rapidly; after finding out the target object, we calculated out the 3-D information of the target object; then PaPeRo decides if continue moving, or adjust the direction, or stop tracking.In our research, we succeed to search out the target object in any direction and tracking automatically.