Tennis Image Recognition System Based on FPGA

• Main Authors: ZI-Shong Chen, 陳子祥
• Other Authors: Ruei-Chang Lu
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/84519156699258645206

碩士 === 國立宜蘭大學 === 電子工程學系碩士班 === 104 === In recent years, the design of robotic tracking systems has become more and more diversified. For example, calculation of paths as well as object shapes, sizes, and colors can be realized through image processing. In this paper, field-programmable gate arrays (FPGAs) are proposed to function as the master controller along with visual image processing and servo motor. This designed system is engineered with a motor module for image recognition of specific target objects and image tracking control. Color-based object tracking remains the focus of research on the design of image tracking systems. In previous studies, multiple design methods for tracking systems have been employed. However, color-based object tracking method that is adopted in most existing studies is designed to recognize target objects from the background that has a fairly high contrast compared to that of the target object. In the present study, the region of interest (ROI) of a tennis image is filtered through an algorithm developed based on the relationship among the RGB value attributes of the image; the tennis can also be recognized from the picture using the algorithm concerned even when the tennis is placed in a scene with lower contrast with the color of the tennis. Moreover, this designed algorithm simplifies post-processing, thereby alleviating the platform processing load. Finally, the recognized tennis imaging information is transferred to the servo motor to simulate motion tracking.