在 OpenCV 中检测纹理？
我正在尝试使用 OpenCV 检测纹理。纹理类似于画笔上的画笔，因此在图像上它会有很多很多的小线条在一起。我尝试使用霍夫线来区分纹理和其他东西，但效果不太好，因为检测到了太多其他误报。除此之外，我有关于使用模板匹配和快速傅里叶变换的想法，但我还没有尝试测试或实现它们。
那么，其他人是否知道可能的方法来做到这一点？也许使用其他线检测器或边缘检测器？或者这会带来太多误报吗？
该纹理应该能够在杂乱的场景中被检测到，并且这样做的算法应该相对较快，因为如果可能的话，我希望实时跟踪它。抱歉，无法发布我想要的纹理样本（重复次数太少），但如果您确实需要查看它的外观，您可以简单地搜索画笔/画笔纹理。但如果您以前见过画笔，那么应该很明显我在谈论哪个部分（有画笔的部分）。
预先感谢，非常感谢。
I'm trying to detect a texture using OpenCV. The texture would be similar to that of a brush on a paintbrush, so on an image it would have many many little lines together. I've tried using Hough Lines to be able to distinguish the texture from other things, but it hasn't been working out too well, as too many other false positives are detected. Other than that, I've had ideas about using Template Matching as well as Fast-Fourier Transforms, but I haven't tried testing or implementing them.
So, would anyone else have any idea of a possible method to do this? Maybe use some other line detector or an edge detector? Or would that bring up too many false positives?
This texture should be able to be detected in a cluttered scene and the algorithm for doing so should be relatively fast, since I want it to be tracked in real-time if possible. Sorry for not being able to post a sample of the texture I want (too less rep l0l), but you can simply search up paintbrush/paintbrush texture if you really need to see what it looks like. But if you've seen a paintbrush before, it should be pretty obvious which part I'm talking about (the part with the brush).
Thanks in advance, really appreciate it.
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