Study on Distance Measurement and Three-Dimensional Modeling Based on Binocular Vision

• Main Authors: Ting-wei Huang, 黃庭偉
• Other Authors: Kuo-yang Tu
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/12905255406406277531

碩士 === 國立高雄第一科技大學 === 電機工程研究所碩士班 === 103 === In recent years, application and development of Intelligent Robot technology is getting more and more widely, including industry, life and competitions that be hold by all districts in the world. In special, FIRA international competition on Intelligent Humanoid Robotics is a contest including Imaging systems, algorithm and hardware implementation. According to all of the above, this thesis is mainly to investigate the application of binocular vision system and image binarization algorithm for distance measurement and three-dimensional modeling in Intelligent Robot Systems. First of all, this thesis is going to make the correction of image distortion for the camera of imaging systems. Eliminate image distortion due to the lens, then probes three kinds of different experiments and applications of image binarization of image processing; however, turn image information into useful resources to supply subsequent use by making use of binarization. Through after the process of above, we could use binocular vision system to get the distance between target and camera to carry out target distance measurement. Secondly, via the image division process to separate the object and background in order to extract feature points from object image to progress the flowing three-dimensional modeling. Last but not least, final results of the experiments will compare the three threshold values of image binarization and each of their pros and cons; as for binocular vision ranging is to put the target at 100 different positions to estimate the object distance by binocular vision algorithm, afterwards, collect the date to calculate the tolerance and the better range of use to present in tabular form; moreover, image ranging is kind of the important thing within intelligent humanoid robotics. Thus, this research results can been implemented for intelligent robots, set the visual system to obtain a more stable and accurate distance measurement capability as the main goal for the purpose that can obtain a better play in the competitions.