Fix motion bugs and add colorscale keyword in plot_image

[pvillacorta]

This PR solves some motion-related bugs that appeared in previous PRs.
Additionally, I added a colorscale keyword for the plot_image function (Let me know if I should include this in another PR or take advantage and include it in this one).

[cncastillo]

I think it's fine, is it ready to merge?

[pvillacorta]

I also took advantage of this PR to add a method definition for the multiplication of an Array of Grads by a rotation Matrix (36ab6c6). This was necessary because rotating an array of gradients (in x, y, and z) with different durations caused all durations to become the same:

julia> G1=Grad(3e-3,1e-3,0)
Grad(3.0 mT, 1.0 ms)

julia> G2=Grad(1.5e-3,2e-3,0)
Grad(1.5 mT, 2.0 ms)

julia> G3=Grad(4.5e-3,0.5e-3,0)
Grad(4.5 mT, 0.5 ms)

julia> g = [G1; G2; G3;;]
3×1 Matrix{Grad}:
 Grad(3.0 mT, 1.0 ms)
 Grad(1.5 mT, 2.0 ms)
 Grad(4.5 mT, 0.5 ms)

julia> R = [0. 0. 1.; 0. -1. 0.; 1. 0. 0.]
3×3 Matrix{Float64}:
 0.0   0.0  1.0
 0.0  -1.0  0.0
 1.0   0.0  0.0

julia> R * g
3×1 Matrix{Grad}:
 Grad(4.5 mT, 2.0 ms)
 Grad(-1.5 mT, 2.0 ms)
 Grad(3.0 mT, 2.0 ms)

I do not know if this is the expected behaviour, but I can assume it is not:

*(x::Sequence, A::Matrix{Float64}) = Sequence(A*x.GR, x.RF, x.ADC, x.DUR, x.DEF) #TODO: change this, Rotation fo waveforms is broken
*(A::Matrix{Float64}, x::Sequence) = Sequence(A*x.GR, x.RF, x.ADC, x.DUR, x.DEF) #TODO: change this, Rotation fo waveforms is broken

Now only the gradient amplitudes rotate, while the durations keep the same (I am not sure if this is correct for all cases):

julia> R * g
3×1 Matrix{Grad}:
 Grad(4.5 mT, 1.0 ms)
 Grad(-1.5 mT, 2.0 ms)
 Grad(3.0 mT, 0.5 ms)

Please let me know if this change is correct, or tell me how you would do it and whether you would include it in this PR as well.
Thank you

[cncastillo]

I don't think the gradient rotation is correct. Do you really need it to be like that?

[pvillacorta]

I have needed this when creating a gre sequence and wanting to acquire different planes with it. In the initial basic sequence, which acquires the XY plane slice, there is a block with gradients of different durations. When rotating the sequence to acquire, for example, the YZ plane, all the gradients of the block end up having the same duration. I think this is an important use case.