mirror of https://github.com/opencv/opencv.git
Open Source Computer Vision Library
https://opencv.org/
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
98 lines
3.2 KiB
98 lines
3.2 KiB
15 years ago
|
#!/usr/bin/env python
|
||
|
|
|
||
|
|
# 2009-01-16, Xavier Delacour <xavier.delacour@gmail.com>
|
||
|
|
|
||
|
|
import unittest
|
||
|
|
from numpy import *;
|
||
|
|
from numpy.linalg import *;
|
||
|
|
import sys;
|
||
|
|
|
||
|
|
import cvtestutils
|
||
|
|
from cv import *;
|
||
|
|
from adaptors import *;
|
||
|
|
|
||
|
|
def transform(H,x):
|
||
|
|
x1 = H * asmatrix(r_[x[0],x[1],1]).transpose()
|
||
|
|
x1 = asarray(x1).flatten()
|
||
|
|
return r_[x1[0]/x1[2],x1[1]/x1[2]]
|
||
|
|
|
||
|
|
class homography_test(unittest.TestCase):
|
||
|
|
|
||
|
|
def test_ransac_identity(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
result,H = cvFindHomography(pts1, pts1, CV_RANSAC, 1e-5);
|
||
|
|
assert(result and all(abs(Ipl2NumPy(H) - eye(3)) < 1e-5));
|
||
|
|
|
||
|
|
def test_ransac_0_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_RANSAC, 1e-5);
|
||
|
|
assert(result and all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_ransac_30_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
pts2[0:30] = random.rand(30,2)
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_RANSAC, 1e-5);
|
||
|
|
assert(result and all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_ransac_70_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
pts2[0:70] = random.rand(70,2)
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_RANSAC, 1e-5);
|
||
|
|
assert(result and all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_ransac_90_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
pts2[0:90] = random.rand(90,2)
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_RANSAC, 1e-5);
|
||
|
|
assert(not result or not all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_lmeds_identity(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
result,H = cvFindHomography(pts1, pts1, CV_LMEDS);
|
||
|
|
assert(result and all(abs(Ipl2NumPy(H) - eye(3)) < 1e-5));
|
||
|
|
|
||
|
|
def test_lmeds_0_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_LMEDS);
|
||
|
|
assert(result and all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_lmeds_30_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = [transform(H1,x) for x in pts1]
|
||
|
|
pts2[0:30] = random.rand(30,2)
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_LMEDS);
|
||
|
|
assert(result and all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def test_lmeds_70_outliers(self):
|
||
|
|
pts1 = random.rand(100,2);
|
||
|
|
H1 = asmatrix(random.rand(3,3));
|
||
|
|
H1 = H1 / H1[2,2]
|
||
|
|
pts2 = vstack([transform(H1,x) for x in pts1])
|
||
|
|
pts2[0:70] = random.rand(70,2)
|
||
|
|
result,H = cvFindHomography(pts1, pts2, CV_LMEDS);
|
||
|
|
assert(not result or not all(abs(H1-H)<1e-5))
|
||
|
|
|
||
|
|
def suite():
|
||
|
|
return unittest.TestLoader().loadTestsFromTestCase(homography_test)
|
||
|
|
|
||
|
|
if __name__ == '__main__':
|
||
|
|
unittest.TextTestRunner(verbosity=2).run(suite())
|
||
|
|
|