Image optical depth incorrect for large velocity gradients

[ThomasCeulemans]

Internally, we assume the trapezoidal rule to be apporximately correct for this imaging method (only valid in case of low velocity gradients); however, we have already switched over the other computations to integrate over the profile function in case of large velocity gradients. This should not be too difficult.