After loading the points from our cameras, we compute the homography vector $\vec{h}$ by solving for a system of equations in which $A\vec{h}=\vec{0}$.
[{'point_id': 1,
'img1_x': 3095.138648488441,
'img1_y': 2982.513633669236,
'img2_x': 923.3420865441612,
'img2_y': 2969.9943687018385},
{'point_id': 2,
'img1_x': 3226.5909306461176,
'img1_y': 2744.647599288679,
'img2_x': 1061.0540011855364,
'img2_y': 2713.3494368701845},
{'point_id': 3,
'img1_x': 3445.678067575578,
'img1_y': 2732.128334321281,
'img2_x': 1255.102608180201,
'img2_y': 2644.4934795494964},
{'point_id': 4,
'img1_x': 3514.5340248962657,
'img1_y': 2563.1182572614116,
'img2_x': 1336.4778304682868,
'img2_y': 2481.743034973326},
{'point_id': 5,
'img1_x': 3358.0432128037937,
'img1_y': 2513.0411973918203,
'img2_x': 1211.2851807943089,
'img2_y': 2444.1852400711327},
{'point_id': 6,
'img1_x': 3746.140426793124,
'img1_y': 2857.3209839952588,
'img2_x': 1524.2668049792528,
'img2_y': 2707.0898043864854},
{'point_id': 7,
'img1_x': 3965.227563722585,
'img1_y': 2857.3209839952588,
'img2_x': 1687.0172495554234,
'img2_y': 2663.272377000593},
{'point_id': 8,
'img1_x': 3821.2560165975106,
'img1_y': 2494.262299940724,
'img2_x': 1586.8631298162418,
'img2_y': 2350.2907528156497},
{'point_id': 9,
'img1_x': 2462.915767634855,
'img1_y': 2337.771487848252,
'img2_x': 203.48435091879037,
'img2_y': 2450.444872554832},
{'point_id': 10,
'img1_x': 2581.8487848251334,
'img1_y': 2225.0981031416723,
'img2_x': 359.97516301126234,
'img2_y': 2287.6944279786608},
{'point_id': 11,
'img1_x': 2588.1084173088325,
'img1_y': 2394.108180201542,
'img2_x': 366.23479549496096,
'img2_y': 2488.002667457025},
{'point_id': 12,
'img1_x': 2588.1084173088325,
'img1_y': 2594.4164196799056,
'img2_x': 359.97516301126234,
'img2_y': 2700.830171902787}]
Thus, we get our 3x3 homography matrix as follows:
array([[ 2.24239286e+00, -1.12331295e-01, -4.90250481e+03],
[ 4.78055032e-01, 1.73964822e+00, -1.13386336e+03],
[ 3.38357384e-04, -6.44486290e-05, 1.00000000e+00]])
Now, we move to warping the images with our given homography matrix. We will use inverse warping, first forwad warping to get the corners of our warped image, then iterating through each pixel location in the warped image and using interpolation to find the correct color value.