表面檢測演算法於Android平台之設計與實作

• Main Authors: Ling-Yu Tsai, 蔡伶鈺
• Other Authors: Chuin-Mu Wang
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/56111555217618720467

碩士 === 國立勤益科技大學 === 資訊工程系 === 104 === With the rapid progress of science and technology of the time, visual images have been from a flat image into 2D, 3D, and 4 d, into a cube of space, and make it become more stereoscopic and vivid. Augmented reality (AR) is the integration of digital information with live video or the user's environment in real time. Basically, AR recognize objects or images in the realistic environment through the lens and blends new information into it. As the result, the technology functions by enhancing one’s current perception of reality. Because convenience as well as software and hardware ability of smart phone promotes, the related application pattern of smart phone is also more and more extensive. Among them to Augmented Reality and image processing related technique applies program (mobile application calls the apps). For example IKEA electronics catalog app, by using augmented reality technique, the user, to drag the virtual image of furniture articles in the catalog into image, and can adjust furniture direction, position and size by oneself. Let the user in advance preview with this image and whether valuation suits, again consideration whether purchases the furniture. This thesis proposes implementation of surface detection algorithm on the Android platform. In the surface detection algorithms, are used to detect features to detect the image of keypoints. For promoting to speed of matching, using BRIEF pattern to obtain the binary descriptors. It can reduces computation of descriptor and promote the matching the descriptors of speed. And then use the RANSAC to remove the false match, and keep the correct match. It can promote the exactitude for matching and save memory and computation. Using the Lucas-Kanade algorithm to initialize the relative position of the feature point, the feature of the image and want to match point in the projected image and immediate use grid lines to show the changes in action. The surface detection algorithm to be applied to very various: The industrial aspect is applied to the machine detect the object in of appearance condition. The smart phone can be applied to pictures, QR codes to detect and then aim at trademarks, OR codes to aim at the different need and different processing action.