Question

我尝试比较ORB和BRISK在图片之间匹配功能的效果。我已将以下函数写入每种方法选择的RANSAC特征匹配：


def match_image_pair(img1, img2, method="ORB"):

    detector = None
    if method == "ORB":
        detector = cv2.ORB()
    elif method == "BRISK":
        t = 30
        octa = 1
        pS = 1.0
        detector = cv2.BRISK(thresh=t, octaves=octa, patternScale=pS)

    if detector == None:
        print "Dodgy method"
        return

    kp_orb, des_orb = detector.detectAndCompute(img1,None)
    src_orb = [o.pt for o in kp_orb]

    kp_orb_test, des_orb_test = detector.detectAndCompute(img2,None)
    src_orb_test = [o.pt for o in kp_orb_test]

    bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
    matches = bf.match(des_orb,des_orb_test)
    print len(matches)
    print len(src_orb), len(src_orb_test)
    M, mask = cv2.findHomography(np.array(src_orb), np.array(src_orb_test), cv2.RANSAC,3.0)
    matchesMask = mask.ravel().tolist()

    matches = [m for m,mask in zip(matches,matchesMask) if mask]

    return drawMatches(img1, kp_orb, img2, kp_orb_test, matches, match_meth=method)

如果method == "ORB"，此函数可以正常工作，但method == "BRISK"时出现异常：


1328
3650 1890
OpenCV Error: Assertion failed (npoints >= 0 && points2.checkVector(2) == npoint
s && points1.type() == points2.type()) in cv::findHomography, file ..\..\..\..\o
pencv\modules\calib3d\src\fundam.cpp, line 1074
Traceback (most recent call last):
  File "match_with_ORB.py", line 134, in 
    match_image_pair(img, img_test, method="BRISK")
  File "match_with_ORB.py", line 114, in match_image_pair
    M, mask = cv2.findHomography(np.array(src_orb), np.array(src_orb_test), cv2.
RANSAC,3.0)
cv2.error: ..\..\..\..\opencv\modules\calib3d\src\fundam.cpp:1074: error: (-215)
 npoints >= 0 && points2.checkVector(2) == npoints && points1.type() == points2.
type() in function cv::findHomography

我可以看到有1328个匹配，并且src_orb和src_orb_test列表分别有3650和1890个元素，因此我无法看到会导致异常的原因I＆＃39;我看到了。

opencv2：findHomography RANSAC使用ORB工作，但使用BRISK提供异常

0 个答案: