| Summary: | 碩士 === 華梵大學 === 電子工程學系碩士班 === 102 === 3C products have many applications in the home life, such as cleaning robot. It can do the work of sweeping the floor for us. Automatical charging is an important function for a cleaning robot. Currently, cleaning robots on the market use the approaches of maze algorithms or wireless communication technology to achieve indoor positioning and looking position of power adaptor. In the first step, they use the way of RFID, Wi-Fi, ZigBee, or other forms of wireless to guide their cleaning robots finding the position of power adaptor, and then use an infrared technology to align their cleaning robots with charging holes for charging. These methods must be used with special power adaptors and put fixed ways of power adaptor to complete the charging alignment action. In order to improve this shortcoming, a new method using image recognition to achieve automatical looking position of power adaptor and aligned movement is proposed in this thesis. In the first step, the proposed method of the thesis is to use rectangular geometric image recognition to find the location of power adaptor. By using experimental methods and statistical analysis, a curve of cleaning robot distance relationship with a rectangular is derived in this thesis. It can be used to find the distance between cleaning robot and power adaptor. After finding the power adaptor, cleaning robot is aligned with the power adaptor by using imaging angle recognition. By using experiments for a number of different scenarios, it is shown that the proposed method can be used in cleaning robot to find the correct position of power adaptor and aligned power adaptor for charging.
Keywords: Cleaning Robot, Automatical Charging, Image Recognition, Indoor Positioning, Wireless Communication, Imaging Angle Recognition, Distance Curve Diagram, Geometric Image Recognition.
|
|---|
