In this chapter,we will explore the quaternions representation of rotations as an alternative to rotation matrices.
在本章节,我们将来一起探究学习作为旋转矩阵替代品的四元数。
Our main use for quaternions will be to aid us in interpolating between orientations in natural way.This will be useful
for animating objects as they fly through space. If we are not planning on interpolating rotations for our application,then this representation my not be necessary.
使用四元数的主要目是以一种自然的方式进行方向间的插值,当动画中的物体在空间移动时,这是非常有用的。当然对于某个特殊的应用,如果你不想进行插值,那么本部分并不是必须的。
7.1 Interpolation
我们假设有一个目标坐标空间在 time=0 时,其旋转变换O0=WR0 ,在time=1 其旋转变换为O1=WR1,R1,R2为4x4的旋转矩阵,假设我们希望找到一系列旋转空间, 能够表示
到
的自然旋转。
p1 :desired object 怎么翻译:
One idea would be to define The problem is that,under matrix linear interpolation,each vector simply moves along a straight line, See Figure7.1. In this case,the intermediate
are not rotation matrices: this is clear unaceptable. Moreover,(especially in 3D),it is not very easy to remove this squashing by somehow finding a true rotation matrix that approximates this interpolated matrix.
一种想法是定义一个在矩阵的线性插值下,这个带来的问题是,每个向量是沿着一条直线进行移动。插值矩阵并不是我们想要的旋转矩阵,这肯定是不能接受的。而且,尤其是在3D空间,找到一个真正的代表旋转的插值矩阵并不是那么容易。
Another idea, not worth pursuing too deely is to somehow to factor both and
into 3,so-called Euler angles(see section 2.5). These three scalar values could each be linearly interpolated using
The interpolated Euler angles could then be used to generate intermediate rotations, it turns out that when interpolating Euler angles the sequences of interpolated rotation is not particularly natural in any physical or geometric way. For examples,it is not invariant to basis changes(define below).
对于上面的描述不太明白,感觉欧拉角应该能得到正确的插值效果,尤其是本段的倒数第二句,没有完全理解
what we would really like to do is to first create a single transition matrix This matrix can,as any rotaion matrix,be thought of as rotation of some
degrees about some axis
, as in Equation(2.5) .Let us image that we had an operation
which give us a rotation about
by
degrees instead. Then it would be a natural to difine
more and more? 这种方式可以构造出 旋转矩阵的球插值效果。
困难的部分是计算出一个能将矩阵转变成轴/角的旋转形式。而四元素就是这样一种形式,而不需要这样的一种转换。
另外使用四元素,我们同样能够像操作矩阵一样操作四元素。
p1, 这里factor 不知道如何对应到文, keep track of ?
7.1.1 周期性
旋转矩阵同样具有一定的周期性。在进行旋转插值时,旋转合适的最小区间是必须要考虑的。
7.1.2 不变性
费解:natural things
7.1 总结:该部分暂时没有理解
7.2 The Reprensentation (表示法)
7.2 总结:讲述了一些四元素的基本概念:能完全理解
7.3 Operations
7.6 So far,we have discussed how quaternions are useful for representing rotations,but have ignored
translation.Now we will discuss how to use quaternions in concert with translation vecors to represent rigid body transformation.
A rigid body transformation,or RBT,can describled by composing a translation and arotation.