如何为图像中的对象提供ID [英] how to give objects in image an ID

问题描述

我正在处理图像处理操作，因为我有建筑物的图像，我正在检测这些物体，并为每个人提供一个ID以便稍后处理，所以我怎样才能给每个人一个ID ？

i''m working on image processing operation as I''ve image of buildings and i''m working on detecting this objects and give each one an ID to deal with later , so how can i give each one an ID ?

推荐答案

ID是一个简单的部分，只需从列表中取下一个数字或使用递增计数器。

困难的部分从一开始就有用地从图像创建对象。我想你正在做边缘检测然后将边缘转换为向量，比较正常的向量找到平行线和计算交叉点或某种交叉点的平均交叉积，以尝试在图像中建立1个或多个消失点？我只是在这里猜测，但如果你已经掌握了所有这些，我就不会过分担心ID'，直到你处理动态图像并试图在多个帧上跟踪同一个物体;那么身份证变得很重要，工作也变得更加困难。如果那就是你需要认真的专家帮助和大量的阅读。我祝你好，但通过视频流进行对象跟踪不在我的联盟中。

The ID is the easy part, just take the next number off a list or use an incrementing counter.
The hard part is usefully creating the objects from an image in the first place. I guess you''re doing edge detection and then coverting the edges to vectors, comparing the normals of the vectors to find parallel lines and computing the intersections or some kind of averaged cross product of the intersections to try and establish 1 or more vanishing points in your image? Im just guessing here but if you''ve already got all that on the go I wouldn''y worry overly about ID''s until you''re dealing with moving images and trying to track the same object across multiple frames; then the ID becomes important and the job a lot harder. If that''s were you''re at you''re going to need to serious specialist help and a lot of reading. I wish you well but object tracking through a video stream is out of my league.

这篇关于如何为图像中的对象提供ID的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！