如何用矩形边界框构造R树(STR方法)?
如何基于矩形边界框排序平铺递归方法并单独搜索来构造R树? 总共 12 个边界框,其中包含 (minx, miny) & (maxx,maxy)
我不需要删除或更新任何节点。
How to construct R Tree based on Sort Tile Recursion Method of rectangle bounding boxes along with searching alone?
Totally 12 bounding boxes with (minx, miny) & (maxx , maxy)
I don't need to delete or update any nodes.
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
检查原始论文...你使用的意思是。
事实上,点数据也会出现这个问题:在第一个批量加载步骤之后,您可能拥有的页面数量超出了根节点的容纳范围,因此您需要批量加载另一个级别。
对于点数据,STR批量加载的页面不会重叠;在更高的水平和对于矩形,他们可能会这样做。 STR 的流行很大程度上源于点数据。
PS 总共有 12 个边界框,使用 R 树对我来说听起来有点大材小用。我将页面大小设置为明显大于 12 个条目。好吧,当你的树由单个节点组成时,它并不是真正的树......
Check the original paper... you use the mean.
In fact this problem arises with point data, too: after a first bulk loading step you may have more pages than fit the root node, so you need to bulk-load another level.
For point data, STR bulk loaded pages will not overlap; at higher levels and for rectangles they will likely do. The popularity of STR probably stems a lot from point data.
P.S. with a total of 12 bounding boxes, using an R-Tree sounds like overkill to me. I'd set the page size clearly larger than 12 entries. And well, when your tree consists of a single node, it isn't really a tree ...