I have a 2D array of Coordinates in Numpy.

My goal is to attempt to find the corners (as if it were a square). So the :

Top left: smallest x, highest y Top right: largest x, largest y bottom left: smallest x, smallest y bottom right: largest x, smallest y

Obviously each of these pairs need to consider the other values.

I was trying to take the min and max depending on the row:

        BottomLeft = np.min(np.min(hull, axis=1), axis=0)

However, this does not keep the pair of values together. It would have to be something like the smallest possible X values, and out of those, the smallest y value. Or something along these lines. I am assuming there is efficient way to do this with numpy?

Here is an example of data:

    [[[260 156]][[248 176]][[235 197]][[233 199]][[192 199]][[174 197]][[160 171]][[150 151]][[154 149]][[156 149]][[260 151]]]

Thanks!

Per the discussion above, this assumes that one of the pairs with the smallest x value will also correspond to the smallest y value. So you can first find the minimum x-value:

# Some sample data
d = np.array([[3, 1, 4, 1, 5],[8, 0, 4, 2, 3]])
# smallest value in the first row which, I assume, is your x-values
xm = np.min(d[0, :])

Then you can get the subset of values that have that minimum x value like so:

d[:, d[0,:] == 1]

So you can get the min of them via:

np.min(d[1, d[0,:] == 1])