SciPy's non-linear least square

Question

Dear All.

I tried to do bundle adjustment by python. So I'm test non-linear least square module. Then I wrote code like below. I want to get right Pmat represents camera projection matrix for three cameras. But I have an error,"ValueError: object too deep for desired array".

Anyone who can give clue to solve this issue?

Regards, Jinho Yoo.

from math import* from numpy import *

import pylab as p from scipy.optimize
import leastsq  

Projected_x = \ mat([[ -69.69  , 255.3825,    1.    ],
        [ -69.69  ,  224.6175,    1.    ],
        [-110.71  ,  224.6175,    1.    ],
        [-110.71  ,  255.3825,    1.    ],
        [ 709.69  ,  224.6175,    1.    ],
        [ 709.69  ,  255.3825,    1.    ],
        [ 750.71  ,  255.3825,    1.    ],
        [ 750.71  ,  224.6175,    1.    ]])

Projected_x = Projected_x.transpose()

Pmat = \ mat(   [[  5.79746167e+02,  0.00000000e+00,   3.20000000e+02,    0.00000000e+00],
        [  0.00000000e+00,   4.34809625e+02,   2.40000000e+02, 0.00000000e+00],
        [  0.00000000e+00,   0.00000000e+00,   1.00000000e+00, 0.00000000e+00] ]  )

reconst_X = \ mat([[-0.95238194, -0.58146697,  0.61506506,  0.00539229],
        [-0.99566105, -0.76178453,  0.72451719,  0.00502341],
        [-1.15401215, -0.81736486,  0.79417098,  0.00546999],
        [-1.11073304, -0.6370473 ,  0.68471885,  0.00583888],
        [ 2.71283058,  2.34190758, -1.80448545, -0.00612243],
        [ 2.7561097 ,  2.52222514, -1.91393758, -0.00575354],
        [ 2.9144608 ,  2.57780547, -1.98359137, -0.00620013],
        [ 2.87118168,  2.39748791, -1.87413925, -0.00656901]])

def residuals(p, y, x):
    err = y - p*x.transpose()

    err = err * err.transpose()

    return err

p0 = Pmat  

plsq = leastsq(residuals, p0, args=(Projected_x, reconst_X  )  )

print plsq[0]

Answer 1

my first guess: leastsq doesn't like matrices,

use arrays and np.dot, or convert np.asarray(err) before the return, And maybe convert p to matrix inside your residual function.

Mixing matrices and arrays can be a pain to keep track of.

Answer 2

A couple of small things:

1. use np.array if you can
2. Do not import *

I have changed the code to use np.array to demonstrate what user333700 means. Also I convert the projection matrix into a 12 dimensional vector since most optimizer expects your variable to optimize in vector form.

The error that you will get running the edited code below is TypeError: Improper input parameters. I believe that this is because you are trying to perform linear least square to find 12 parameters but you only have 8 constraints.

import numpy as np

import pylab as p
from scipy.optimize import leastsq

Projected_x = np.array([[ -69.69  , 255.3825,    1.    ],
        [ -69.69  ,  224.6175,    1.    ],
        [-110.71  ,  224.6175,    1.    ],
        [-110.71  ,  255.3825,    1.    ],
        [ 709.69  ,  224.6175,    1.    ],
        [ 709.69  ,  255.3825,    1.    ],
        [ 750.71  ,  255.3825,    1.    ],
        [ 750.71  ,  224.6175,    1.    ]])

Projected_x = Projected_x.transpose()

Pmat = np.array(   [  5.79746167e+02,  0.00000000e+00,   3.20000000e+02,    0.00000000e+00,
          0.00000000e+00,   4.34809625e+02,   2.40000000e+02, 0.00000000e+00,
          0.00000000e+00,   0.00000000e+00,   1.00000000e+00, 0.00000000e+00]   )

reconst_X = np.array([[-0.95238194, -0.58146697,  0.61506506,  0.00539229],
        [-0.99566105, -0.76178453,  0.72451719,  0.00502341],
        [-1.15401215, -0.81736486,  0.79417098,  0.00546999],
        [-1.11073304, -0.6370473 ,  0.68471885,  0.00583888],
        [ 2.71283058,  2.34190758, -1.80448545, -0.00612243],
        [ 2.7561097 ,  2.52222514, -1.91393758, -0.00575354],
        [ 2.9144608 ,  2.57780547, -1.98359137, -0.00620013],
        [ 2.87118168,  2.39748791, -1.87413925, -0.00656901]])

def residuals(p, y, x):
    err = y - np.dot(p.reshape(3,4),x.T)

    print p

    return np.sum(err**2, axis=0)

p0 = Pmat

plsq = leastsq(residuals, p0, args=(Projected_x, reconst_X  )  )

print plsq[0]