如何计算屏幕上所有物体/形状的总面积
我正在尝试计算所有身体或形状在屏幕上占据的总面积。即，如果我有两个相互相交的圆 A 和 B，我想计算 A 并 B 覆盖的面积（在屏幕上）。
我一直在阅读花栗鼠文档并查看花栗鼠API 的方法，我可能使用，但我还没有找到可以直接使用的东西。
我发现的唯一两种可能有用的方法是这两种： pointQueryFirst:layers:group: 和 segmentQueryFirstFrom:to:layers:group:
我的想法是：
- 使用第一种方法（pointQueryFirst）遍历屏幕上的所有点并调用此方法。如果一个点没有形状，则将其累积到一个变量中。然后将该变量值除以屏幕面积即可得到空闲屏幕的百分比。
- 或者使用第二种方法（segmentQueryFirstFrom），创建一个递归算法，将屏幕分为两半，并在每一半上运行查询，如果任何一半包含形状，则将该区域分成两半并检查这些区域是否包含任何形状，等等上...
但我预计在使用它们时，整体性能会受到影响。我可以使用其他解决方案吗？我还没有找到的另一种方法？非常感谢任何帮助。
I'm trying to calculate the total area all bodies or shapes are occupying on the screen. I.e. if I have 2 circles, A and B, that intersect each other, I want to calculate the area that A union B covers (on the screen).
I've been reading through the chipmunk documentation and looked in the chipmunk API for a method that I might use, but I haven't found anything that I can use directly.
The only two methods I found, that might be useful, are these two: pointQueryFirst:layers:group: and segmentQueryFirstFrom:to:layers:group:
The way I was thinking was to:
- Use the first method (pointQueryFirst) to go through all points on the screen and call this method. If a point doesn't have a shape in them, then accumulate it to a variable. Then divide that variable value with the area of the screen to get percentage of the screen that is free.
- Or use the second method (segmentQueryFirstFrom), create an recursive algorithm that divides the screen in half and run the query on each half, if any half contains a shape, then divide that area into halves and check if those contains any shapes, and so on...
But I expect that in using them, the overall performance will suffer. Is there another solution that I can use? Another method that I haven't found? Any help is greatly appreciated.
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）
Chipmunk 并不特别能够帮助你解决这个问题。你提到的方法可行，但速度慢得离谱。
我想我会做一个很好的老式遮挡查询。将形状渲染到纹理或某种离屏缓冲区中，然后计算像素。
Chipmunk isn't particularly going to be able to help you with that. The methods you mentioned will work, but be ridiculously slow.
I think I would do a good old fashioned occlusion query. Render the shapes into a texture or some sort of offscreen buffer and then count the pixels.