Collisions during calibration

[romainbrette]

I believe during one of my tests the objective and pipette collided. This could happen in two cases:

• the manipulator advances (X) and the microscope moves down simultaneously, but too fast.
• the pipette goes up before the microscope goes up. To solve this, several possible solutions:
• make sure motor speeds match on different axes (possibly complicated)
• never move the pipette up, always down (then you need to know what is up and down)
• always move the microscope down at the end, after the pipette has moved.

[romainbrette]

This requires to know what is "up" and "down". For this, we use the fact that pipettes are always inclined downwards. When we move along the X axis, the pipette moves on the image. We can detect whether it advances on the image, for example using template matching together with the cardinal position (there's this Aurélien's code). So can we associate a direction to the axis (ie +1 = advance). Then we can estimate the focal Z move; from this, we deduce the direction of the Z axis (ie, +1 = down). From that, we can guess also the direction of the Z axis of the manipulator.

Alternatively, we require the user to provide the direction of microscope Z for going down.

[romainbrette]

Done