互联网旅游信息挖掘和展示技术研究
摘要
本文首先通过Flickr提供的数据下载接口得到带有经纬度信息图像以及其对应元数据(如上传用户ID、上传时间、拍摄时间等)、地理标签(经纬度信息)和文本标签等信息。在这些信息基础上,提出了结合K-means聚类和AP聚类的图像聚类算法,以获得在空间位置上相近的图像集合。
然后,给定包含空间上相近图像所组合成集合,对集合中每幅图像提取其SIFT(Scale-invariant feature transform)特征,并利用KD树存储从每幅图像中所提取的SIFT特征,然后通过BBF(Best-Bin First)算法搜索该集合中任意两幅图像之间相似特征点,构建视觉相似图像集合。为从该集合中获取代表性标签来标注该集合图像,本文在TF-IDF(term frequency-inverse document frequency)的基础上结合实际提出了WTF-IDF-UF,最后利用Naive Scan Methods去除那些随机性标签。最终,获得每一类在空间相近、在视觉上相似图像集合所对应热门地标名称、地理位置、热门度和代表性图像等属性信息。
为了对这些空间相近、视觉相似的图像集合及其标签进行存储,本文通过全球四叉树按照显示级别将地理空间划分为不同大小正方形,对图像集合进行编码,然后将地理空间上任一点经纬度值映射为所在级别的四叉树编码,以进行索引。
最后,本文利用Google Maps API和AJAX技术来直观、动态的展示上述信息。
In this thesis, API of Flickr is implemented to obtain images and their corresponding metadata (such as the image ID, uploading time, the time the images taken photograph), the geo tags and annotated textual tags.
After getting abovementioned information, we use k-means and AP algorithms together to group similar images within neighboring spatial distances. In order to find visually similar images within each group, BBF(Best-Bin First) algorithm is conducted to align each images by SIFT local features. Then, WTF-IDF-UF is conducted to obtain informative tags for each group which contain visually similar images with neighbored distance, and the Naive Scan Methods are used to remove random irrelevant tags. At last, we can get the name, location, popularity of the hot landmarks as well as the representative images for each image group.
In order to better store each image group and their corresponding information, a global quadtree is introduced in this thesis. The basic idea of global quadtree is to segment the map into different squares and map any location in the map to a quadcode at any levels.
at last we use Google maps API and AJAX to show the information intuitively and dynamically.
然后,给定包含空间上相近图像所组合成集合,对集合中每幅图像提取其SIFT(Scale-invariant feature transform)特征,并利用KD树存储从每幅图像中所提取的SIFT特征,然后通过BBF(Best-Bin First)算法搜索该集合中任意两幅图像之间相似特征点,构建视觉相似图像集合。为从该集合中获取代表性标签来标注该集合图像,本文在TF-IDF(term frequency-inverse document frequency)的基础上结合实际提出了WTF-IDF-UF,最后利用Naive Scan Methods去除那些随机性标签。最终,获得每一类在空间相近、在视觉上相似图像集合所对应热门地标名称、地理位置、热门度和代表性图像等属性信息。
为了对这些空间相近、视觉相似的图像集合及其标签进行存储,本文通过全球四叉树按照显示级别将地理空间划分为不同大小正方形,对图像集合进行编码,然后将地理空间上任一点经纬度值映射为所在级别的四叉树编码,以进行索引。
最后,本文利用Google Maps API和AJAX技术来直观、动态的展示上述信息。
In this thesis, API of Flickr is implemented to obtain images and their corresponding metadata (such as the image ID, uploading time, the time the images taken photograph), the geo tags and annotated textual tags.
After getting abovementioned information, we use k-means and AP algorithms together to group similar images within neighboring spatial distances. In order to find visually similar images within each group, BBF(Best-Bin First) algorithm is conducted to align each images by SIFT local features. Then, WTF-IDF-UF is conducted to obtain informative tags for each group which contain visually similar images with neighbored distance, and the Naive Scan Methods are used to remove random irrelevant tags. At last, we can get the name, location, popularity of the hot landmarks as well as the representative images for each image group.
In order to better store each image group and their corresponding information, a global quadtree is introduced in this thesis. The basic idea of global quadtree is to segment the map into different squares and map any location in the map to a quadcode at any levels.
at last we use Google maps API and AJAX to show the information intuitively and dynamically.
引文
[1]Flickr, http://www.flickr.com[OL].
[2]Tongqiang Guo, Lei Tian, Shoucun Ji, Binwei Wu. Tourism Service System Based on Ubiquitous Location Aware Infrastructure [C].2nd IEEE International Conference on Ubi-media Computing, Taiwan,2009:12.
[3]David G. Lowe. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision,2004,60(2):91-110.
[4]J. Beis, D. Lowe. Shape indexing using approximate nearest-neighbour search inhighdimensional spaces[C]. In Conference on Computer Vision and PatternRecognitio PuertoRico,1997:1000-1006.
[5]Lyndon Kennedy, Mor Naaman. Generating Diverse and Representative ImageSearch Results for Landmarks[C]. WWW 2008, Beijing, China,2008.
[6]YanTao Zheng, Ming Zhao, Yang Song, Hartwig Adam, Ulrich Buddemeier. Tour the World:building a web-scale landmark recognition engine[C]. Proceedings ofInternational Conference on Computer Vision and Pattern Recognition Miami, Florida, U.S.A,2009.
[7]Ian Simon, Noah Snavely, Steven M. Seitz. Scene Summarization for OnlineImage Collections[C]. IEEE 11th International Conference on In Computer Vision,2007.
[8]Tye Rattenbury, Nathaniel Good, Mor Naaman. Towards Automatic Extraction of Event and Place Semantics from Flickr Tags[C]. SIGIR'07, Amsterdam, The Netherlands,2007.
[9]Till Quack, Bastian Leibe, Luc Van Gool. World-scale Mining of Objects and Events from Community Photo Collections[C]. CIVR'08, Niagara Falls, Ontario, Canada,2008.
[10]Shane Ahern, Mor Naaman, Rahul Nair, Jeannie Yang. World Explorer: Visualizing Aggregate Data from Unstructured Text in Geo-Referenced Collections[J]. JCDL'07, June 17-22,2007, Vancouver, British Columbia, Canada.
[11]Ting Liu, Andrew W. Moore, Alexander Gray and Ke Yang. An Investigation of Practical Approximate Nearest Neighbor Algorithms[N]. MIT Press 2004: 825-832.
[12]Yu shi Jing, Shumeet Baluja, Henry Rowley. Canonical Image Selection from the Web[C]. CIVR'07, Amsterdam, The Netherlands,2007.
[13]Sham Kakade, Greg Shakhnarovich. Metric trees for fast NN search[C]. CMSC 35900 Large Scale Learning,2009.
[14]David G. Lowe. Object Recognition from Local Scale-Invariant Features[C]. Proc. of the International Conference on Computer Vision, Corfu,1999.
[15]Shuxiang Li, Murray H.Loew. The quadcode and its arithmetic [J]. Communications of the ACM,1987,30(7):621-626.
[16]Kanth V Kothuri, Siva Ravada, Daniel Abugov. Quadtree and R-tree indexes in oracle spatial:a comparison using GIS data[C]. Proceedings of the 2008 ACM SIGMOD International Conference on Management of Data. New York, 2008:546-557.
[17]HungYi Lin. Using B+-trees for processing of line segments in large spatial databases[J]. Journal of Intelligent Information Systems,2008,31(1):35-52.
[18]Yuhang Ding, Dayan Zhuang, Kejun Wang. A Study of Hand Recognition Method [C]. Proceedings of the IEEE International Conference on Mechatronics & Automation Niagara Falls, Canada,2005.
[19]Google Maps API, http://code.google.com/intl/zh-CN/apis/maps/index.html[OL].
[20]H.Steinhaus.Sur la division des corps materiels en parties(in French)[J]. Bulletin L'Academie Polonaise des Science,1956,4(12):801-804.
[21]B.J. Frey, D.Dueck. Clustering by passing messages between data points[J]. Science,2007,312:972-976.
[22]冀寿存.基于位置相关的旅游信息推送服务关键技术研究[D].硕士学位论文,浙江大学,2009.
[23]蔡胜渊.基于几何约束的图像配准研究[D].硕士学位论文,浙江大学,2010.
[24]百度百科.http://baike.baidu.com/view/739359.html [OL].
[25]百度百科.http://baike.baidu.com/view/1228847.htm[OL].
[26]百度文库.http://wenku.baidu.com/view/7152424769eae009581bec5a.html[OL].
[2]Tongqiang Guo, Lei Tian, Shoucun Ji, Binwei Wu. Tourism Service System Based on Ubiquitous Location Aware Infrastructure [C].2nd IEEE International Conference on Ubi-media Computing, Taiwan,2009:12.
[3]David G. Lowe. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision,2004,60(2):91-110.
[4]J. Beis, D. Lowe. Shape indexing using approximate nearest-neighbour search inhighdimensional spaces[C]. In Conference on Computer Vision and PatternRecognitio PuertoRico,1997:1000-1006.
[5]Lyndon Kennedy, Mor Naaman. Generating Diverse and Representative ImageSearch Results for Landmarks[C]. WWW 2008, Beijing, China,2008.
[6]YanTao Zheng, Ming Zhao, Yang Song, Hartwig Adam, Ulrich Buddemeier. Tour the World:building a web-scale landmark recognition engine[C]. Proceedings ofInternational Conference on Computer Vision and Pattern Recognition Miami, Florida, U.S.A,2009.
[7]Ian Simon, Noah Snavely, Steven M. Seitz. Scene Summarization for OnlineImage Collections[C]. IEEE 11th International Conference on In Computer Vision,2007.
[8]Tye Rattenbury, Nathaniel Good, Mor Naaman. Towards Automatic Extraction of Event and Place Semantics from Flickr Tags[C]. SIGIR'07, Amsterdam, The Netherlands,2007.
[9]Till Quack, Bastian Leibe, Luc Van Gool. World-scale Mining of Objects and Events from Community Photo Collections[C]. CIVR'08, Niagara Falls, Ontario, Canada,2008.
[10]Shane Ahern, Mor Naaman, Rahul Nair, Jeannie Yang. World Explorer: Visualizing Aggregate Data from Unstructured Text in Geo-Referenced Collections[J]. JCDL'07, June 17-22,2007, Vancouver, British Columbia, Canada.
[11]Ting Liu, Andrew W. Moore, Alexander Gray and Ke Yang. An Investigation of Practical Approximate Nearest Neighbor Algorithms[N]. MIT Press 2004: 825-832.
[12]Yu shi Jing, Shumeet Baluja, Henry Rowley. Canonical Image Selection from the Web[C]. CIVR'07, Amsterdam, The Netherlands,2007.
[13]Sham Kakade, Greg Shakhnarovich. Metric trees for fast NN search[C]. CMSC 35900 Large Scale Learning,2009.
[14]David G. Lowe. Object Recognition from Local Scale-Invariant Features[C]. Proc. of the International Conference on Computer Vision, Corfu,1999.
[15]Shuxiang Li, Murray H.Loew. The quadcode and its arithmetic [J]. Communications of the ACM,1987,30(7):621-626.
[16]Kanth V Kothuri, Siva Ravada, Daniel Abugov. Quadtree and R-tree indexes in oracle spatial:a comparison using GIS data[C]. Proceedings of the 2008 ACM SIGMOD International Conference on Management of Data. New York, 2008:546-557.
[17]HungYi Lin. Using B+-trees for processing of line segments in large spatial databases[J]. Journal of Intelligent Information Systems,2008,31(1):35-52.
[18]Yuhang Ding, Dayan Zhuang, Kejun Wang. A Study of Hand Recognition Method [C]. Proceedings of the IEEE International Conference on Mechatronics & Automation Niagara Falls, Canada,2005.
[19]Google Maps API, http://code.google.com/intl/zh-CN/apis/maps/index.html[OL].
[20]H.Steinhaus.Sur la division des corps materiels en parties(in French)[J]. Bulletin L'Academie Polonaise des Science,1956,4(12):801-804.
[21]B.J. Frey, D.Dueck. Clustering by passing messages between data points[J]. Science,2007,312:972-976.
[22]冀寿存.基于位置相关的旅游信息推送服务关键技术研究[D].硕士学位论文,浙江大学,2009.
[23]蔡胜渊.基于几何约束的图像配准研究[D].硕士学位论文,浙江大学,2010.
[24]百度百科.http://baike.baidu.com/view/739359.html [OL].
[25]百度百科.http://baike.baidu.com/view/1228847.htm[OL].
[26]百度文库.http://wenku.baidu.com/view/7152424769eae009581bec5a.html[OL].