Integration System of Computer Vision Positioning for Single Camera

• Main Authors: Lin,Pinyang, 林品洋
• Other Authors: Yeh,Shuming
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/03055463280271292142

碩士 === 醒吾技術學院 === 資訊科技應用研究所 === 99 === Positioning is the mainly point in our research. To complete more and more mission, a robot should be keep improve with more and more function, including the basic action "move". In order to avoid the collision, the best way is setting the positioning system to ensure the robot would not leave the designated route, and real-time display the exactly location of the robot. Inertial positioning is the most common way was adopted by robot, but while moving more and more distance, the error would be accumulated slowly then result in collision occurred, or even get lost. If pick high accuracy positioning system like LIDAR or laser, would greatly enhance the robot set-up cost. Therefore, in order to reduce set-up cost and reach to high positioning accuracy, we will propose an improvement of positioning system with our research. We adopt computer vision positioning system, by switching to use appearance detecting positioning and infrared ray source positioning in different lights conditions, and also run ZigBee to do a regional positioning to strengthen effect of computer vision positioning.Setting up a webcam in the center of ceiling, then installed ZigBee in the targeting place. At the same time, also equip infrared light launch and special images on the carrier that need to be positioned in advance, include ZigBee. Using the built-in positioning function (measure the energy of electric) of ZigBee to calculate the approximate distance between each ZigBee and ZigBee by strength of radio wave was sent by itself. (Error about 2 meters) After determining the respective regions, operate the camera to observe. During day time, we could determine the position by lattice characteristic that be formed by 5x5 black and white squares. In the night, we should put a film filter in front of camera to help us easy to check the emissive position of infrared rays, and take the vision for positioning.In the future, we could put the infrared rays launch and special images on the carrier which need to be positioned, and using ZigBee to assist the computer vision positioning system at the same time. By this way, we could achieve the original goal that we set up at the first about lower cost and more accurate positioning system. Besides, the Zigbee which set on the carrier also can be a channel to transmit the simple signal. And the Zigbee which set on the specific location not only assist to position, but also can be a router of others ZigBee to assist them to transmit some information, or increase the date transmitted area.