How to make computers understand length, breadth and height given a 3d object?

Question

Suppose, I am building an A.I Robot and i want that robot to bring to me that cuboidal shape object whose length is say 10 cm(in 1st case out of (40cm, 19cm, 10cm)), each time it brings the desired object i place the object back there with its bottom(base) resting on its different side so how will it distinguish b/w the length, breadth and height of the object? For example, the next three images are of the same cuboid having bases on different sides image1 image2 image3.Thinking in pseudo code, my opinion is that for height(side which is ground up or,) comparing it with a virtual vertical unit scale will do the work, but i am unable to come up with how to find the difference b/w lengths and breadths. How can i figure out the difference b/w the length and breadth of this 3d cuboid? Also just by simply placing the cuboid with the base having a diffrent side interchanges the values of l,b,h completely most of the time(or atleast two of the values interchange every time) And now what to do when i need to generalize the situation for many different cuboids with different length, breadth and height? (skip this part answering if the answer is similar to the previous part).